Parallel CURE using MPI in C

The project is a parallel implementation of the CURE clustering alorithm using MPI and C.

The main idea is to distribute data to multiple processors and then sample them using chernoff bound to determine sample size, where a local CURE will be executed followed by outlier elimination.

It will then accumulate cluster representatives to 'master' processor and do the second pass of CURE with the clusters created by every processor.

The data sould be read from a file and be of k dimensions. The distance metric is euclidian distance. Outliers are clusters with a small density (prefferably by dividing number of cluster objects by cluster area).

A final outlier elimination stage and data membership is required, as well as computation timing for io and clustering.

Please do not hesitate to ask for any further details.

Pemrograman C

ID Proyek: #6226177

Tentang proyek