latent semantic analysis via the singular value decomposition (for dummies)

introduction

example 1: two clear features

example 2: two less clear features

example 3: two less clear features (revisited)

using real data

using real data (with document length normalisation)

conclusions

get the code from github