Linear Regression for Car Price Prediction

A Linear regression model to predict the price of a car based on its mileage. The model is implemented based on gradient descent algorithm.

Prerequisites

Python3
Numpy
Matplotlib

Getting started

Clone the repository.

Running the programs

Predicting car price

To start the first program, run the following command:

python3 predict.py

If the model is already trained, the program will use the saved theta0 and theta1 values to predict the car's price for a given mileage. If the program hasn't been trained yet, the program will use the default values of theta0 and theta1 to make the prediction(both set to 0).

Training the model

To run the second program, run the following command:

python3 train.py --path=path/to/dataset.csv --display=True

This program will read the specified dataset file (specified using --path argument)and perform a linear regression on the data. After the linear regression is completed, theta0 and theta1 will be saved for use in the first program. If the --display is set to True, the program will display a graph showing the data distribution and the fit line of the model.

visualizing the results

The visualized results below gives a clear illustation of the model's performance:

The fit-line graph displays the relationship between the real values and the predicted values, represented by data points and the best fit-line.
The loss over iterations graph shows the reduction of the loss function value with each iteration of the training process until it reaches the optimal solution.

Note

The learning rate and number of iterations can be changed in the train.py file.