pyclustering 0.8.1

pyclustering 0.8.1 library is collection of clustering algorithms, oscillatory networks, neural networks, etc.

GENERAL CHANGES:

• Implemented feature to use specific metric for distance calculation in K-Means algorithm (pyclustering.cluster.kmeans, ccore.clst.kmeans). See: https://github.com/annoviko/pyclustering/issues/434

• Implemented BANG-clustering algorithm with result visualizer (pyclustering.cluster.bang). See: https://github.com/annoviko/pyclustering/issues/424

• Implemented feature to use specific metric for distance calculation in K-Medians algorithm (pyclustering.cluster.kmedians, ccore.clst.kmedians). See: https://github.com/annoviko/pyclustering/issues/429

• Supported new type of input data for K-Medoids - distance matrix (pyclustering.cluster.kmedoids, ccore.clst.kmedoids). See: https://github.com/annoviko/pyclustering/issues/418

• Implemented TTSAS algorithm (pyclustering.cluster.ttsas, ccore.clst.ttsas). See: https://github.com/annoviko/pyclustering/issues/398

• Implemented MBSAS algorithm (pyclustering.cluster.mbsas, ccore.clst.mbsas). See: https://github.com/annoviko/pyclustering/issues/398

• Implemented BSAS algorithm (pyclustering.cluster.bsas, ccore.clst.bsas). See: https://github.com/annoviko/pyclustering/issues/398

• Implemented feature to use specific metric for distance calculation in K-Medoids algorithm (pyclustering.cluster.kmedoids, ccore.clst.kmedoids). See: https://github.com/annoviko/pyclustering/issues/417

• Implemented distance metric collection (pyclustering.utils.metric, ccore.utils.metric). See: no reference.

• Supported new type of input data for OPTICS - distance matrix (pyclustering.cluster.optics, ccore.clst.optics). See: https://github.com/annoviko/pyclustering/issues/412

• Supported new type of input data for DBSCAN - distance matrix (pyclustering.cluster.dbscan, ccore.clst.dbscan). See: no reference.

• Implemented K-Means observer and visualizer to visualize and animate clustering results (pyclustering.cluster.kmeans, ccore.clst.kmeans). See: no reference.

CORRECTED MAJOR BUGS:

• Bug with out of range in K-Medians (pyclustering.cluster.kmedians, ccore.clst.kmedians). See: https://github.com/annoviko/pyclustering/issues/428

• Bug with fast linking in PCNN (python implementation only) that wasn't used despite the corresponding option (pyclustering.nnet.pcnn). See: https://github.com/annoviko/pyclustering/issues/419