Scribd
Upload
35 views2 pages

Ieee Research Paper

1) The document discusses using predictive analytics and machine learning models to predict electronic gadget prices based on specifications. 2) Various models were tested on training data and random forest was found to have the highest accuracy of 88.7%. 3) Random forest works by creating many decision trees on random subsets of data and averaging the results. 4) When the model was tested on test data, the predicted prices closely matched actual prices with an error of only 0.15%, validating the model.

Uploaded by

anishdayani2
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
35 views2 pages

Ieee Research Paper

1) The document discusses using predictive analytics and machine learning models to predict electronic gadget prices based on specifications. 2) Various models were tested on training data and random forest was found to have the highest accuracy of 88.7%. 3) Random forest works by creating many decision trees on random subsets of data and averaging the results. 4) When the model was tested on test data, the predicted prices closely matched actual prices with an error of only 0.15%, validating the model.

Uploaded by

anishdayani2
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

IV.

Proposed Approach for finding the best suitable model and the testing data
will be used to find the difference from the experimental
Predictive analytics is a branch of advanced analytics that
predictions. After testing each model using the training
makes predictions about future outcomes using historical
data, we obtained the R2 Score (Accuracy) and MAE
data combined with statistical modelling, data mining
(Mean accuracy error) given below in the tabular form.
techniques and machine learning. Companies employ
predictive analytics to find patterns in this data to identify MODEL ACCURACY MAE
risks and opportunities. Predictive analytics is often Linear Regression 80.7% 0.21
associated with big data and data science. Our Project Ridge Regression 81.2% 0.20
“Wise Price” uses predictive analysis to analyse and KNN 80.2% 0.19
predict the prices of electronic gadgets using its Random Forest 88.7% 0.15
specifications. SVM 80.8% 0.20
Fig 2. Table showing the accuracy and Error margin generated by
A. Theoretical model, analysis, and solution each predic on model.

Predictive analysis involves series of steps which need to As per the table given in Fig 2, highest accuracy was
be followed. The first step was collection of data. The generated by Random Forest model. Therefore, Random
data which we collected included the name, specifications Forest model was chosen as the Prediction model for our
(RAM, ROM, SSD/HDD, Screen Resolution, Type, PPI) project.
and price of these laptops. The next step involved
preprocessing of data. Raw data consists of blank Random Forest is a popular machine learning algorithm
columns and inconsistencies. Thus, series of steps such as that belongs to the supervised learning technique. It can
data cleaning and data transformation was applied for be used for both Classification and Regression problems
preprocessing. Further we performed exploratory data in ML. It is based on the concept of ensemble
analysis (EDA) to visualize and summarize the data to learning, which is a process of combining multiple
gain insights into its characteristics. classifiers to solve a complex problem and to improve the
performance of the model. As the name
suggests, "Random Forest is a classifier that contains a
number of decision trees on various subsets of the given
dataset and takes the average to improve the predictive
accuracy of that dataset." Instead of relying on one
decision tree, the random forest takes the prediction from
each tree and based on the majority votes of predictions,
and it predicts the final output.

Fig 1. EDA used for plo ng a bar graph to find which companies
produce budget friendly electronics.

One of the exploratory analyses was performed on the


data to find out the price range at which companies
release their electronic gadgets. Many such analysis were
performed to find the factors which will be helpful in
predicting algorithms for maximum accuracy.
Fig 3. Diagramma c representa on of Random Forest Algorithm
The most important step is the selection of prediction
model to predict the most accurate prices. There are many The Working of the Random Forest Algorithm is quite
models present for prediction such as Linear Regression, intuitive. It is implemented in two phases: The first is to
K-Means algorithm, Logistic regression, Decision Tress, combine N decision trees with building the random
Random Forest and Support Vector Machine (SVM) forest, and the second is to make predictions for each
model. Our pre-processed data was divided into two parts tree created in the first phase.
– training data and testing data. The training data is used
The following steps can be used to demonstrate the Laptop Experimental Actual
working process: Dell 14(Intel i5) ₹ 52,120.7015 ₹ 52,199
Acer Aspire ₹ 27,948.015 ₹ 27,990
Step 1: Pick M data points at random from the training Lite
set. HP Laptop 15s ₹ 28,946.515 ₹ 28,990
MSI GF63 ₹ 64,292.515 ₹ 64,990
Step 2: Create decision trees for your chosen data points ASUS Tuff ₹ 57,903.015 ₹ 57,990
(Subsets). Lenovo IdeaPad ₹ 56,804.795 ₹ 56,990
Apple 2022 ₹ 1,35,786.015 ₹ 1,35,990
Step 3: Each decision tree will produce a result. Analyse Fig 5. Table shows the comparison between Experimental and Actual
it. Prices.

By analysing the results from the above table, one can


Step 4: For classification and regression, accordingly,
the final output is based on Majority Voting or observe that difference or margin of error is nearly 0.15%
Averaging, accordingly. as given in the table in Fig 2. With accuracy of the model
up to 89%, the results obtained are very close to the actual
results and thus satisfactory.
In conclusion, the comparison between the actual and
experimental results has yielded highly satisfactory
outcomes. The consistency and agreement between the
two sets of data validate the accuracy and reliability of
our predictive model. This alignment not only reinforces
our confidence in the predictive capabilities of the model
but also underscores its practical utility in real-world
applications. These results not only affirm the
effectiveness of our approach but also pave the way for
informed decision-making and enhanced problem-
solving in the future.
Fig 4. Working of the Random Forest Prediction Model

Consider a scenario consisting of a dataset containing


several fruits images. And the Random Forest Classifier
is given this dataset. Each decision tree is given a subset
of the dataset to work with. During the training phase,
each decision tree generates a prediction result. The
Random Forest classifier predicts the final decision
based on most outcomes when a new data point appears.

The last step involved rigorously testing the model with


the testing dataset which was bifurcated from our pre-
processed dataset in the earlier stages. This assisted in
fine tuning the parameters in the model such as batch size
and epoch to further increase the accuracy of the model.
B. Comparison of Experimental and Actual Results
The model was deemed satisfactory by further testing and
fine tuning the parameters of the model. To compare the
experimental results with the actual results, a table was
maintained for the above results. For ease of
understanding, prediction was performed on prices of
laptops with various brands and specifications. A total of
7 observations can be seen in the tabular form that
compares the actual price of the laptop with the prices
predicted by the Random Forest Model.

You might also like