Principles of Knowledge‐Based Guidance in Fuzzy Clustering
Witold Pedrycz · 2005
Proposed is taxonomy of knowledge-oriented hints about data to be structured where the domain knowledge involves aspects of given similarity between patterns (known as proximity hints) or uncertainty of labeling (then reflected by entropy hints). The detailed algorithm of proximity-based fuzzy clustering, P-FCM, is discussed in detail discussing the rationale and revealing optimization aspects of the method. The application facet involves WWW exploration (search for similar Web pages).