First experience with theano library.
This script creates a 2D synthetic dataset that corresponds to x^2 with some added noise. It then tries to fit a model to that dataset and plots the results.
This is a very simple example of a Linear Regression method. By adding additional features, it can fit non-linear curves, such as the x^2 curve.
Still needs some improvements:
- Regularization should be added
- The Learning rate should be dynamic
- The algorithm to add new features should be improved