Research Of Fuzzy Clustering Ensemble Approaches

In recent year, ensemble learning has becoming a research focus in the data minging and machine learning. As an important part of ensemble learning, clustering ensemble also has been widely applied to various fields. Performing clustering analysis is one of the important research topics in cancer discovery using gene expression profile, which is crucial in facilitating the successful diagnosis and treatment of cancer.Most of the existing research works adopt single clustering algorithms to perform tumor clustering from bio-molecular data which lack of robustness, stability and accuracy, and few of them considers how to incorporate fuzzy theory together with an optimization process into a consensus clustering framework to improve the performance of clustering analysis. Aiming at covering the above shortcomings, this essay proposed two workable approaches. Firstly, we introduce the fuzzy theory into the cluster ensemble framework for tumor clustering from bio-molecular data, and propose four kinds of hybrid fuzzy cluster ensemble frameworks,link the components in serial way or in concurrent way, and distribute the samples based on the fuzzy membership function to obtain the final result. Secondly, we introduce the random double clustering based fuzzy cluster ensemble framework(RDCFCE), which is designed to improve the performance of RDCCE by integrating the newly proposed fuzzy extension model into the ensemble framework. Further more, adaptive RDCFCE(A-RDCFCE) is proposed to optimize RDCFCE and improve the performance of RDCFCE further by adopting a self-evolutionary process(SEPP) for the parameter set. To measure the performances of the proposed two approaches, we separately utilized UCI machine learning data sets and gene expression data sets. The experiments illustrate that(1) the proposed fuzzy cluster ensemble frameworks work well on real datasets, especially bio-molecular data.(2) The proposed approaches are able to provide more robust, stable and accurate results when compared with the state-of-the-art single clustering algorithms and traditional cluster ensemble approaches.