A Link-Based Fuzzy Clustering Ensemble

A link-based fuzzy cluster ensemble (LBFCE) is proposed to solve the problem that many clustering ensemble methods ignore the underlying information or acquire the underlying information by complex approaches. In the LBFCE, an ensemble information matrix is first built by primarily exploiting the results of fuzzy clustering, this matrix is then transformed into a weighted graph with data relations by appropriate link analysis, and at last a graph partitioning algorithm is employed to get the final clustering results. Experimental results show that the LBFCE algorithm may obtain the underlying information effectively and improve clustering performance.