A ew Method for Clustering Heterogeneous Data: Clustering by Compression

Nowadays, we have to deal with a large quantity of unstructured data, produced by a number of sources. For example, clustering web pages is essential to getting structured information in response to user queries. In this paper, we intend to test the results of a new clustering technique – clustering by compression – when applied to heterogeneous sets of data. The clustering by compression procedure is based on a parameterfree, universal, similarity distance, the normalized compression distance or NCD, computed from the lengths of compressed data files (singly and in pair-wise concatenation). Compression algorithms allow defining a similarity measure based on the degree of common information, whereas clustering methods allow clustering similar data without any previous knowledge. Key-Words: clustering, heterogeneous data, clustering by compression, Normalized Compression Distance (NCD), FScore