Fuzzy Clustering of Heterogeneous Patterns
Witold Pedrycz · 2005
Discussed is clustering of heterogeneous data that are patterns whose features are information granules rather than numeric entities. Those are represented with the aid of fuzzy sets and fuzzy relations. The fundamental design phase concerns the representation of information granules for clustering and representation of results (e.g., prototypes) that leads to several mechanisms of encoding and decoding. Two fundamental approaches to clustering known as parametric and nonparametric techniques are analyzed.