Car_Price_Prediction

A model to predict the price of a car using Data Science and Machine learning in Python.

Training Accuracy : 97.75%
Testing Accuracy : 87.89%

Abstract

Preprocessing

The data is cleaned.
Outliers are removed.
Categorical values are converted to numeric format (Label Encoding).

Fearure Selection:

Correlation of all features with each other is checked.
ExtraTreesRegressor is used to check the importance of all features with respect to target variable.
Features with less importance are dropped.

Split the dataset into train and test dataset.

Build the predictor model

Random Forest Regressor: Training accuracy of 97.75% and Testing accuracy of 87.89%
Linear Regression : Training accuracy of 82.37% and Testing accuracy of 76.54%
Thus, Random Forest Regressor Model was chosen.

Visualization

Some Graphs were presented to show the result:
Histogram
Heatmap
Line Graph