Locality-Sensitive Hashing for High-Dimensional Clustering

Clone the package and create virtualenv by running the appropriate make target:

cd lsh-hdc/
make env

Some original parts specific to LSH were written by Mike Starr. The algorithms were modified for the high-dimensional clustering application on short text content by Eugene Scherba. The lsh_hdc.metrics and other submodules were written by Eugene Scherba. Some methods are borrowed (with tweaks) from the Scikit-Learn project, in which case the docstring for the methods notes this.

The current project starting from version v0.2.0 is under BSD 3-clause license due to inclusion of parts of Scikit-Learn code, which is under the same license. The original code by Mike Starr as well as the current package up until version v0.1.5 is under MIT license.