Sparse Matrix Representations of Feature Space using Eigen3

Co-occurrence analysis - bread and butter metrics and filtering provides the following tools:

samples2cooc - create co-occurrence matrix and feature list from samples
cooc2logl - compute log likelihood of co-occurrences
loglfilter - filter input co-occurrence matrix based on logl threshold
AB-cosine - compare feature vectors in co-occurrence matrix
show_features - human readable feature vectors using co-occurrence matrix and feature list

Also provides some general eigen3 matrix utilities

dumpmat - dump matrix contents to stdout useful for eyeballing computations and debugging

And useful C++ headers

index_bimap.hpp
sparse_matrix.hpp
sparse_matrix_io.hpp