This project objective is to recode a linear regression function in order to predict the price of a car depending on the mileage.
A basic implementation gives us a good evolution of the cost and prediction in approximately 1000 iterations.
Some bonus parts were made during this project :
- data_generator : a data generator script to challenge the performance of the implementation
- regularization : L1 and L2 regularization
- decay rate
- tol : stop the training of the cost does not differ much between 2 iterations
- batch gradient descent (example of cost evolution below)
