Overview

This is a Python implementation of the the linear manifold clustering algorithm developed by Haralick and Harpaz (see LICENSE for full citation).

Module documentation is in docs/build/html/index.html

Examples of usage are in lmclus/lmclus.py and testlm.py, but:

from lmclus import lmclus

D = Dataset (observations*feature vectors) K = maximum dimension of all linear manifolds S = approximate number of clusters Gamma = threshold for deciding if a manifold is a good fit

C, labels = lmclus.LMCLUS(D, K, S, Gamma)

C is a Python dictionary where each cluster (starting at 0) is a key, and each value is a dictionary containing attributes of the cluster.

labels is a list of cluster labels for the observations in D