FUZZY CLUSTERING METHOD USING PRINCIPAL COMPONENT ANALYSIS AND MARKOV CHAIN MONTE CARLOS

摘要

PURPOSE: fuzzy clustering method is used to Markov chain Montecarle be combined to provide to obtain conditional probability distribution, without complicated calculating. ;CONSTITUTION: the quantity of initial group is determined (S110). The hypothesis of dictionary distribution is about initial group. Gibbs sampler algorithms are applied to. A kind of pillar layout density function assumes that. Column probability value is calculated, according to the gibbs sample (S120) for using extraction. There is greatest member's function value to determine about each entity (S130) for the group. ;The 2012 of copyright KIPO submissions