Block Sparse Representations in Modified Fuzzy C-Regression Model Clustering Algorithm for TS Fuzzy Model Identification

A novel objective function based clustering algorithm has been introduced by considering linear functional relation between input-output data and geometrical shape of input data. Noisy data points are counted as a separate class and remaining good data points in the data set are considered as good clusters. This noise clustering concept has been taken into the proposed objective function to obtain the fuzzy partition matrix of product space data. Block orthogonal matching pursuit algorithm is applied to determine the optimal number of rules from the over specified number of rules (clusters). The obtained fuzzy partition matrix is used to determine the premise variable parameters of Takagi-Sugeno (TS) fuzzy model. Once, the premise variable parameters and optimal number of rules (clusters) are identified then formulate the rule construction for identification of linear coefficients of consequence parameters. The effectiveness of the proposed algorithm has been validated on two benchmark models.