Codigo SVM
Uploaded by
DIANA PAZ PINO ABARCACodigo SVM
Uploaded by
DIANA PAZ PINO ABARCA#Kernel Lineal
#SVM1
svm_model1 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=0.01,gamma=0.01)
summary(svm_model1) #realiza 8326 vectores de soporte
#training set predictions
pred_train <-predict(svm_model1,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8914234
#test set predictions
pred_test <-predict(svm_model1,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8907442
#Calculamos métricas
pred_test <-predict(svm_model1,dfval)
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8907442
error_rate <- (fp+fn)/total
error_rate #0.1092558
recall <- tp/(tp+fn)
recall #0.1548507
sensitivity <- tp/positive
sensitivity #0.1548507
especificity <- tn/negative
especificity#0.9897127
precision <- tp/predicted_positive
precision #0.6693548
F1 <- 2/(1/precision+1/recall)
F1 #0.2515152
data.frame(accuracy,precision,recall,F1)
#SVM2
svm_model2 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=0.1,gamma=0.01)
summary(svm_model2) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model2,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model2,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
pred_test <-predict(svm_model2,dfval)
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8967157
recall <- tp/(tp+fn)
recall #0.2677239
sensitivity <- tp/positive
sensitivity #0.2677239
especificity <- tn/negative
especificity #0.9813072
precision <- tp/predicted_positive
precision #0.6582569
F1 <- 2/(1/precision+1/recall)
F1 # 0.3806366
data.frame(accuracy,precision,recall,F1)
#SVM3
svm_model3 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=1,gamma=0.01)
summary(svm_model3) #realiza 8039 vectores de soporte
#training set predictions
pred_train <-predict(svm_model3,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8994968
#test set predictions
pred_test <-predict(svm_model3,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.897158
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.897158
recall <- tp/(tp+fn)
recall #0.2826493
sensitivity <- tp/positive
sensitivity #0.2826493
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6530172
F1 <- 2/(1/precision+1/recall)
F1 # 0.3945313
data.frame(accuracy,precision,recall,F1)
#SVM4
svm_model4 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=10,gamma=0.01)
summary(svm_model4) #realiza 8030 vectores de soporte
#training set predictions
pred_train <-predict(svm_model4,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8998009
#test set predictions
pred_test <-predict(svm_model4,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8974898
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.3976608
data.frame(accuracy,precision,recall,F1)
#SVM5
svm_model5 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=100,gamma=0.01)
summary(svm_model5) #realiza 8 vectores de soporte
#training set predictions
pred_train <-predict(svm_model5,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model5,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
#SVM6
svm_model6 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=0.01,gamma=0.1)
summary(svm_model6) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model6,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model6,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
#SVM7
svm_model7 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=0.1,gamma=0.1)
summary(svm_model7) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model7,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model7,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM8
svm_model8 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=1,gamma=0.1)
summary(svm_model8) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model8,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model8,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM9
svm_model9 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=10,gamma=0.1)
summary(svm_model9) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model9,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model9,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM10
svm_model10 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=100,gamma=0.1)
summary(svm_model10) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model10,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model10,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM11
svm_model11 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=0.01,gamma=1)
summary(svm_model11) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model11,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model11,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM12
svm_model12 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=0.1,gamma=1)
summary(svm_model12) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model12,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model12,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM13
svm_model13 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=1,gamma=1)
summary(svm_model13) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model13,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model13,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM14
svm_model14 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=10,gamma=1)
summary(svm_model14) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model14,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model14,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM15
svm_model15 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=100,gamma=1)
summary(svm_model15) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model15,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model15,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM16
svm_model16 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=0.01,gamma=10)
summary(svm_model16) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model16,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model16,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM17
svm_model17 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=0.1,gamma=10)
summary(svm_model17) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model17,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model17,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM18
svm_model18 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=1,gamma=10)
summary(svm_model18) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model18,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model18,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM19
svm_model19 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=10,gamma=10)
summary(svm_model19) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model19,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model19,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM20
svm_model20 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="linear",cost=100,gamma=10)
summary(svm_model20) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model20,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model20,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#Kernel Radial
#SVM1
svm_model1 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=0.01,gamma=0.01)
summary(svm_model1) #realiza 8326 vectores de soporte
#training set predictions
pred_train <-predict(svm_model1,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8914234
#test set predictions
pred_test <-predict(svm_model1,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8907442
#Calculamos métricas
pred_test <-predict(svm_model1,dfval)
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8907442
error_rate <- (fp+fn)/total
error_rate #0.1092558
recall <- tp/(tp+fn)
recall #0.1548507
sensitivity <- tp/positive
sensitivity #0.1548507
especificity <- tn/negative
especificity#0.9897127
precision <- tp/predicted_positive
precision #0.6693548
F1 <- 2/(1/precision+1/recall)
F1 #0.2515152
data.frame(accuracy,precision,recall,F1)
#SVM2
svm_model2 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=0.1,gamma=0.01)
summary(svm_model2) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model2,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model2,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
pred_test <-predict(svm_model2,dfval)
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8967157
recall <- tp/(tp+fn)
recall #0.2677239
sensitivity <- tp/positive
sensitivity #0.2677239
especificity <- tn/negative
especificity #0.9813072
precision <- tp/predicted_positive
precision #0.6582569
F1 <- 2/(1/precision+1/recall)
F1 # 0.3806366
data.frame(accuracy,precision,recall,F1)
#SVM3
svm_model3 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=1,gamma=0.01)
summary(svm_model3) #realiza 8039 vectores de soporte
#training set predictions
pred_train <-predict(svm_model3,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8994968
#test set predictions
pred_test <-predict(svm_model3,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.897158
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.897158
recall <- tp/(tp+fn)
recall #0.2826493
sensitivity <- tp/positive
sensitivity #0.2826493
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6530172
F1 <- 2/(1/precision+1/recall)
F1 # 0.3945313
data.frame(accuracy,precision,recall,F1)
#SVM4
svm_model4 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=10,gamma=0.01)
summary(svm_model4) #realiza 8030 vectores de soporte
#training set predictions
pred_train <-predict(svm_model4,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8998009
#test set predictions
pred_test <-predict(svm_model4,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8974898
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.3976608
data.frame(accuracy,precision,recall,F1)
#SVM5
svm_model5 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=100,gamma=0.01)
summary(svm_model5) #realiza 8 vectores de soporte
#training set predictions
pred_train <-predict(svm_model5,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model5,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
#SVM6
svm_model6 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=0.01,gamma=0.1)
summary(svm_model6) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model6,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model6,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
#SVM7
svm_model7 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=0.1,gamma=0.1)
summary(svm_model7) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model7,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model7,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM8
svm_model8 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=1,gamma=0.1)
summary(svm_model8) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model8,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model8,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM9
svm_model9 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=10,gamma=0.1)
summary(svm_model9) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model9,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model9,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM10
svm_model10 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=100,gamma=0.1)
summary(svm_model10) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model10,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model10,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM11
svm_model11 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=0.01,gamma=1)
summary(svm_model11) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model11,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model11,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM12
svm_model12 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=0.1,gamma=1)
summary(svm_model12) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model12,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model12,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM13
svm_model13 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=1,gamma=1)
summary(svm_model13) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model13,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model13,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM14
svm_model14 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=10,gamma=1)
summary(svm_model14) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model14,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model14,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM15
svm_model15 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=100,gamma=1)
summary(svm_model15) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model15,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model15,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM16
svm_model16 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=0.01,gamma=10)
summary(svm_model16) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model16,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model16,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM17
svm_model17 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=0.1,gamma=10)
summary(svm_model17) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model17,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model17,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM18
svm_model18 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=1,gamma=10)
summary(svm_model18) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model18,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model18,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM19
svm_model19 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=10,gamma=10)
summary(svm_model19) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model19,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model19,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM20
svm_model20 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="radial",cost=100,gamma=10)
summary(svm_model20) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model20,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model20,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#Kernel Polynomial
#SVM1
svm_model1 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=0.01,gamma=0.01)
summary(svm_model1) #realiza 8326 vectores de soporte
#training set predictions
pred_train <-predict(svm_model1,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8914234
#test set predictions
pred_test <-predict(svm_model1,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8907442
#Calculamos métricas
pred_test <-predict(svm_model1,dfval)
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8907442
error_rate <- (fp+fn)/total
error_rate #0.1092558
recall <- tp/(tp+fn)
recall #0.1548507
sensitivity <- tp/positive
sensitivity #0.1548507
especificity <- tn/negative
especificity#0.9897127
precision <- tp/predicted_positive
precision #0.6693548
F1 <- 2/(1/precision+1/recall)
F1 #0.2515152
data.frame(accuracy,precision,recall,F1)
#SVM2
svm_model2 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=0.1,gamma=0.01)
summary(svm_model2) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model2,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model2,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
pred_test <-predict(svm_model2,dfval)
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8967157
recall <- tp/(tp+fn)
recall #0.2677239
sensitivity <- tp/positive
sensitivity #0.2677239
especificity <- tn/negative
especificity #0.9813072
precision <- tp/predicted_positive
precision #0.6582569
F1 <- 2/(1/precision+1/recall)
F1 # 0.3806366
data.frame(accuracy,precision,recall,F1)
#SVM3
svm_model3 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=1,gamma=0.01)
summary(svm_model3) #realiza 8039 vectores de soporte
#training set predictions
pred_train <-predict(svm_model3,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8994968
#test set predictions
pred_test <-predict(svm_model3,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.897158
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.897158
recall <- tp/(tp+fn)
recall #0.2826493
sensitivity <- tp/positive
sensitivity #0.2826493
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6530172
F1 <- 2/(1/precision+1/recall)
F1 # 0.3945313
data.frame(accuracy,precision,recall,F1)
#SVM4
svm_model4 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=10,gamma=0.01)
summary(svm_model4) #realiza 8030 vectores de soporte
#training set predictions
pred_train <-predict(svm_model4,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8998009
#test set predictions
pred_test <-predict(svm_model4,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8974898
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.3976608
data.frame(accuracy,precision,recall,F1)
#SVM5
svm_model5 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=100,gamma=0.01)
summary(svm_model5) #realiza 8 vectores de soporte
#training set predictions
pred_train <-predict(svm_model5,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model5,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
#SVM6
svm_model6 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=0.01,gamma=0.1)
summary(svm_model6) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model6,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model6,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
#SVM7
svm_model7 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=0.1,gamma=0.1)
summary(svm_model7) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model7,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model7,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM8
svm_model8 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=1,gamma=0.1)
summary(svm_model8) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model8,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model8,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM9
svm_model9 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=10,gamma=0.1)
summary(svm_model9) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model9,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model9,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM10
svm_model10 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=100,gamma=0.1)
summary(svm_model10) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model10,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model10,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM11
svm_model11 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=0.01,gamma=1)
summary(svm_model11) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model11,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model11,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM12
svm_model12 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=0.1,gamma=1)
summary(svm_model12) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model12,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model12,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM13
svm_model13 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=1,gamma=1)
summary(svm_model13) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model13,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model13,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM14
svm_model14 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=10,gamma=1)
summary(svm_model14) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model14,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model14,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM15
svm_model15 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=100,gamma=1)
summary(svm_model15) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model15,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model15,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM16
svm_model16 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=0.01,gamma=10)
summary(svm_model16) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model16,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model16,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM17
svm_model17 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=0.1,gamma=10)
summary(svm_model17) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model17,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model17,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM18
svm_model18 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=1,gamma=10)
summary(svm_model18) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model18,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model18,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM19
svm_model19 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=10,gamma=10)
summary(svm_model19) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model19,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model19,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM20
svm_model20 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="polynomial",cost=100,gamma=10)
summary(svm_model20) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model20,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model20,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#Kernel Gaussiana
#SVM1
svm_model1 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=0.01,gamma=0.01)
summary(svm_model1) #realiza 8326 vectores de soporte
#training set predictions
pred_train <-predict(svm_model1,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8914234
#test set predictions
pred_test <-predict(svm_model1,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8907442
#Calculamos métricas
pred_test <-predict(svm_model1,dfval)
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8907442
error_rate <- (fp+fn)/total
error_rate #0.1092558
recall <- tp/(tp+fn)
recall #0.1548507
sensitivity <- tp/positive
sensitivity #0.1548507
especificity <- tn/negative
especificity#0.9897127
precision <- tp/predicted_positive
precision #0.6693548
F1 <- 2/(1/precision+1/recall)
F1 #0.2515152
data.frame(accuracy,precision,recall,F1)
#SVM2
svm_model2 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=0.1,gamma=0.01)
summary(svm_model2) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model2,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model2,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
pred_test <-predict(svm_model2,dfval)
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8967157
recall <- tp/(tp+fn)
recall #0.2677239
sensitivity <- tp/positive
sensitivity #0.2677239
especificity <- tn/negative
especificity #0.9813072
precision <- tp/predicted_positive
precision #0.6582569
F1 <- 2/(1/precision+1/recall)
F1 # 0.3806366
data.frame(accuracy,precision,recall,F1)
#SVM3
svm_model3 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=1,gamma=0.01)
summary(svm_model3) #realiza 8039 vectores de soporte
#training set predictions
pred_train <-predict(svm_model3,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8994968
#test set predictions
pred_test <-predict(svm_model3,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.897158
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.897158
recall <- tp/(tp+fn)
recall #0.2826493
sensitivity <- tp/positive
sensitivity #0.2826493
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6530172
F1 <- 2/(1/precision+1/recall)
F1 # 0.3945313
data.frame(accuracy,precision,recall,F1)
#SVM4
svm_model4 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=10,gamma=0.01)
summary(svm_model4) #realiza 8030 vectores de soporte
#training set predictions
pred_train <-predict(svm_model4,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8998009
#test set predictions
pred_test <-predict(svm_model4,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8974898
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.3976608
data.frame(accuracy,precision,recall,F1)
#SVM5
svm_model5 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=100,gamma=0.01)
summary(svm_model5) #realiza 8 vectores de soporte
#training set predictions
pred_train <-predict(svm_model5,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model5,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
#SVM6
svm_model6 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=0.01,gamma=0.1)
summary(svm_model6) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model6,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model6,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
#SVM7
svm_model7 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=0.1,gamma=0.1)
summary(svm_model7) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model7,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model7,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM8
svm_model8 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=1,gamma=0.1)
summary(svm_model8) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model8,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model8,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM9
svm_model9 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=10,gamma=0.1)
summary(svm_model9) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model9,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model9,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM10
svm_model10 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=100,gamma=0.1)
summary(svm_model10) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model10,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model10,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM11
svm_model11 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=0.01,gamma=1)
summary(svm_model11) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model11,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model11,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM12
svm_model12 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=0.1,gamma=1)
summary(svm_model12) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model12,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model12,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM13
svm_model13 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=1,gamma=1)
summary(svm_model13) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model13,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model13,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM14
svm_model14 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=10,gamma=1)
summary(svm_model14) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model14,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model14,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM15
svm_model15 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=100,gamma=1)
summary(svm_model15) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model15,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model15,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM16
svm_model16 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=0.01,gamma=10)
summary(svm_model16) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model16,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model16,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM17
svm_model17 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=0.1,gamma=10)
summary(svm_model17) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model17,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model17,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM18
svm_model18 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=1,gamma=10)
summary(svm_model18) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model18,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model18,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM19
svm_model19 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=10,gamma=10)
summary(svm_model19) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model19,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model19,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
#SVM20
svm_model20 <- svm(Deposito_a_plazo~ ., data=dftrain, method="C-classification",
kernel="gaussian",cost=100,gamma=10)
summary(svm_model20) #realiza 8.104 vectores de soporte
#training set predictions
pred_train <-predict(svm_model20,dftrain)
mean(pred_train==dftrain$Deposito_a_plazo) #0.8976167
#test set predictions
pred_test <-predict(svm_model20,dfval)
mean(pred_test==dfval$Deposito_a_plazo) #0.8967157
#Calculamos métricas
performance_data<-data.frame(observed=dfval$Deposito_a_plazo,
predicted= pred_test)
positive <- sum(performance_data$observed=="1")
negative <- sum(performance_data$observed=="0")
predicted_positive <- sum(performance_data$predicted=="1")
predicted_negative <- sum(performance_data$predicted=="0")
total <- nrow(performance_data)
data.frame(positive, negative,predicted_positive,predicted_negative)
tp<-sum(performance_data$observed=="1" & performance_data$predicted=="1")
tn<-sum(performance_data$observed=="0" & performance_data$predicted=="0")
fp<-sum(performance_data$observed=="0" & performance_data$predicted=="1")
fn<-sum(performance_data$observed=="1" & performance_data$predicted=="0")
data.frame(tp,tn,fp,fn)
accuracy <- (tp+tn)/total
accuracy #0.8974898
recall <- tp/(tp+fn)
recall # 0.2854478
sensitivity <- tp/positive
sensitivity #0.2854478
especificity <- tn/negative
especificity #0.9798018
precision <- tp/predicted_positive
precision #0.6552463
F1 <- 2/(1/precision+1/recall)
F1 # 0.2854478
data.frame(accuracy,precision,recall,F1)
data.frame(accuracy,precision,recall,F1)
You might also like
- Support Vector Machine With Multiple Classes100% (1)Support Vector Machine With Multiple Classes5 pages
- UNIT-II-Support Vector Machine AlgorithmNo ratings yetUNIT-II-Support Vector Machine Algorithm13 pages
- SVM Classifier For Multi Class Image ClassificationNo ratings yetSVM Classifier For Multi Class Image Classification2 pages
- SVM, Neural Network and Random Forest in RNo ratings yetSVM, Neural Network and Random Forest in R45 pages
- Title: Implement Support Vector Machine Classifier: Department of Computer Science and EngineeringNo ratings yetTitle: Implement Support Vector Machine Classifier: Department of Computer Science and Engineering5 pages
- A Practical Guide To Support Vector Classification: I I I N LNo ratings yetA Practical Guide To Support Vector Classification: I I I N L15 pages
- Support Vector Machine: Classification, Regression and Outliers DetectionNo ratings yetSupport Vector Machine: Classification, Regression and Outliers Detection26 pages
- A Practical Guide To Support Vector Classi Cation - Chih-Wei Hsu, Chih-Chung Chang and Chih-Jen LinNo ratings yetA Practical Guide To Support Vector Classi Cation - Chih-Wei Hsu, Chih-Chung Chang and Chih-Jen Lin12 pages
- A Practical Guide To Support Vector Classification: I I I N LNo ratings yetA Practical Guide To Support Vector Classification: I I I N L12 pages
- Financial Time Series Forecasting Using SVMNo ratings yetFinancial Time Series Forecasting Using SVM21 pages
- EIE520 Neural Computation: The Hong Kong Polytechnic UniversityNo ratings yetEIE520 Neural Computation: The Hong Kong Polytechnic University14 pages
- An Introduction Of: Support Vector MachineNo ratings yetAn Introduction Of: Support Vector Machine36 pages
- An Introduction To Support Vector MachinesNo ratings yetAn Introduction To Support Vector Machines22 pages
- An Introduction To Support Vector MachinesNo ratings yetAn Introduction To Support Vector Machines22 pages
- Aim of The Experiment-Software Required - TheoryNo ratings yetAim of The Experiment-Software Required - Theory6 pages
