Implementation of "Collaborative Filtering On a Budget" by Karatzoglou et.
al.

Python class that allows you to perform large scale collaborative filtering
tasks.  Makes use of hash kernels to bound the memory usage of the program,
independent of the feature space, or size of the corpus.

See "Hash Kernels for Structured Data" by Shi et. al. for more details.

Depends on the mmhash python package