Rough Fuzzy -Medoids and Amino Acid Sequence Analysis

The problem with using most pattern recognition algorithms to analyze biological subsequences is that they cannot recognize nonnumerical features such as the biochemical codes of amino acids. Investigating a proper encoding process before modeling the amino acids is then critical. This chapter uses the rough-fuzzy c-medoids to select a minimum set of most informative bio-basis strings. It introduces necessary notions of bio-basis function and the bio-basis string selection methods, and gives a description of hard c-medoids, fuzzy c-medoids, possibilistic c-medoids, and fuzzy-possibilistic c-medoids. The chapter describes the rough-fuzzy c-medoids algorithm, along with the rough c-medoids algorithm, for relational data clustering. It also gives an overview of the application of different c-medoids algorithms for bio-basis string selection, along with the initialization method for the c-medoids algorithm. The chapter presents some quantitative measures to select most informative bio-basis strings, and reports on a few case studies and a comparison among different methods.

fuzzy set theory; pattern clustering; rough set theory