P Y T H O N F O R D ATA Working On Model

SCIENCE Model Choosing Train-Test

Data
C H E AT S H E E T Supervised Learning Estimator:
• Linear Regression:
• Naive Bayes:
>>> from sklearn.naive_bayes import
GaussianNB
Unsupervised Learning Estimator:
• Principal Component Analysis (PCA):
>>> from sklearn.decomposition import
Supervised:
>>> from sklearn.linear_model import >>>new_ lr.fit(X, y)
LinearRegression >>> new_gnb = GaussianNB() PCA
>>> knn.fit(X_train, y_train)
• KNN:
Python Scikit-Learn >>> new_lr =
LinearRegression(normalize=True) >>> from sklearn import neighbors
>>>
>>> new_pca= PCA(n_components=0.95)
• K Means:
>>> from sklearn.cluster import KMeans
>>>new_svc.fit(X_train, y_train)
Unsupervised :
• Support Vector Machine: >>> k_means.fit(X_train)
>>> from sklearn.svm import SVC knn=neighbors.KNeighborsClassifier(n_ne >>> k_means = KMeans(n_clusters=5,
random_state=0) >>> pca_model_fit =
>>> new_svc = SVC(kernel='linear') ighbors=1)
new_pca.fit_transform(X_train)
Introduction
Scikit-learn:“sklearn" is a machine learning library for the Python programming language.
Simple and efficient tool for data mining, Data analysis and Machine Learning. Post-Processing
Importing Convention - import sklearn

Preprocessing Prediction Model Tuning

Supervised: Grid Search: Randomized Parameter Optimization:

>>> y_predict = >>> from sklearn.grid_search import GridSearchCV >>> from sklearn.grid_search import RandomizedSearchCV
Data Loading Train-Test new_svc.predict(np.random.random((3,5))) >>> params = {"n_neighbors": np.arange(1,3), "metric": >>> params = {"n_neighbors": range(1,5), "weights":
• Using NumPy: >>> y_predict = new_lr.predict(X_test) ["euclidean", "cityblock"]} ["uniform", "distance"]}
Data >>> y_predict = knn.predict_proba(X_test) >>> grid = GridSearchCV(estimator=knn, >>> rsearch = RandomizedSearchCV(estimator=knn,
>>>import numpy as np param_grid=params) param_distributions=params, cv=4, n_iter=8, random_state=5)
>>>a=np.array([(1,2,3,4),(7,8,9,10)],dtype=int) >>> grid.fit(X_train, y_train) >>> rsearch.fit(X_train, y_train)
>>>data = np.loadtxt('file_name.csv', >>>from sklearn.model_selection Unsupervised:
>>> y_pred = k_means.predict(X_test) >>> print(grid.best_score_) >>> print(rsearch.best_score_)
delimiter=',') import train_test_split
• Using Pandas:
>>> print(grid.best_estimator_.n_neighbors)

>>>import pandas as pd >>> X_train, X_test, y_train, y_test =

>>>df=pd.read_csv file_name.csv ,header=0) train_test_split(X,y,random_state=0) Evaluate Performance
Classification: Regression: Clustering: Cross-validation:
Data Preparation 1. Confusion Matrix: 1. Mean Absolute Error:
>>> from sklearn.metrics import mean_absolute_error
1. Homogeneity: >>> from
>>> from sklearn.metrics import >>> from sklearn.metrics import sklearn.cross_validation

• Standardization • Normalization
confusion_matrix homogeneity_score import cross_val_score
>>> print(confusion_matrix(y_test, >>> y_true = [3, -0.5, 2] >>> homogeneity_score(y_true, >>>
>>>from sklearn.preprocessing import >>>from sklearn.preprocessing import y_pred)) >>> mean_absolute_error(y_true, y_predict) y_predict) print(cross_val_score(knn,
StandardScaler Normalizer 2. Accuracy Score: 2. Mean Squared Error: 2. V-measure: X_train, y_train, cv=4))
>>>get_names = df.columns >>> knn.score(X_test, y_test) >>> from sklearn.metrics import mean_squared_error >>> from sklearn.metrics import >>>
>>>pd.read_csv("File_name.csv")
>>>scaler = >>> from sklearn.metrics import >>> mean_squared_error(y_test, y_predict) v_measure_score print(cross_val_score(new_
>>>x_array = np.array(df[ Column1 ]
preprocessing.StandardScaler() accuracy_score 3. R² Score : >>> metrics.v_measure_score(y_true, lr, X, y, cv=2))
#Normalize Column1
>>>scaled_df = scaler.fit_transform(df) >>> accuracy_score(y_test, y_pred) >>> from sklearn.metrics import r2_score y_predict)
>>>normalized_X =
>>>scaled_df = >>> r2_score(y_true, y_predict)
preprocessing.normalize([x_array])
pd.DataFrame(scaled_df,
columns=get_names)m
FURTHERMORE:
Python for Data Science Certification Training Course