A Special FCL Clustering and Its Application to Sparse Blind Source Separation
Tong Yu, Yunjie Zhang · Procedia Engineering · 2011
Abstract Fuzzy c-Lines (FCL) algorithm is a linear fuzzy clustering algorithm and is constructed to treat linear dates and capture linear substructures. This paper considers a special FCL clustering algorithm, then based on the special FCL clustering algorithm and generalized inverse of vectors, proposes a new two-step clustering algorithms in order to solve the underdetermined sparse blind signal separation. The proposed algorithm provides a new approach for mixing matrix estimation and source signals separation, and simulation results support the validity of the algorithm.