Multivariate linear regression with gradient descent

this is a Java implementation of the Gradient descent algorithm, applied to Multivariate linear regresion. (Only one feature is provided but we generate more by exponentiation of the one provided: x^2, x^3... x^31). The algorithm is applied the same way if we have more natural features (but it would be harder to represent graphically).

The implementation is based on the following update rules, explained by prof. Andrew Ng at the Stanford-Machine Learning course in Coursera:

In this example you click on the graph to add new points and then use the "fit" button to execute gradient descent and fit a polynomial expression to the data provided.