It has been a while, but finally there is a new note on numpy/scipy recipes ...

Starting on 12 April 2016, I will give a lecture course on deep learning for multimedia pattern recognition. The lecture will cover both theory and numerous practical applications.

The nearest neighbor problem frequently arises in may areas of data science. In this short note, we show how to efficiently tackle it in numpy.

There are several new videos illustrating how k-means clustering and its variants behave in practice.

Data scientists frequently need to compute the eigenvalues and eigenvectors of data covariance matrices. We therefore discuss appropriate ways of how this can be done using numpy/scipy.