An IPython notebook on the linear perceptron

Just mirroring a post from the other blog.

Here is an IPython notebook with an implementation of the linear percepton algorithm.

Details will follow in another post and I give a general idea of what it does in the notebook but here is what the picture looks like.

Assuming that you have a set of points and you want to find a line that separates the points into different categories, this algorithm is one way to do this. It picks an initial line and moves it around the space until it finds a good separator of the set. At the end, the result should look like the example below.


You can also fork that notebook on github.