Including auxiliary information in fuzzy clustering
Paul R. Kersten · 2002
Fuzzy clustering is an effective prerequisite exploratory data analysis (EDA) tool to aid in the design of pattern recognition classifiers. Often, signal-to-noise (SNR) information is available with each data sample. The fuzzy C-Means (FCM) and the fuzzy C-Medians (FCMED) clustering algorithms are extended to include this auxiliary information. These extensions are tested using Slash data where the variation in SNR is assumed to be available to the system. Two fuzzy clustering methods are compared with and without the auxiliary information available.