The first AI project of 42. It aims to introduce us to machine learning.
We have to create a program that predicts the price of a car by using a linear function train with a gradient descent algorithm.
- First, run the training for the prediction to work:
$ python3 train.py- Then, just run the prediction and enter your desired mileage:
$ python3 predict.py
Loading data...
Enter mileage: <number>If your value is incorrect you should get the
Invalid input!error and you will be asked to input the mileage again.
- You now have a predicted price of a car
Price: 8000-
The values are first normalized (scaled between 0 and 1) to be easier to process.
-
Normalized thetas are then calculated via gradient descent
-
A denormalized version of thetas is calculated so that it can be used with normal values.
-
We can now predict the price of a car with an affine function (with thetas), and voila!