作者：
Qian, ZhengXia, HongxiaVocat Coll
Deans Off Elect Informat Technol Bengbu 233000 Anhui Peoples R China Vocat Coll
Sch Software Elect Informat Technol Bengbu 233000 Anhui Peoples R China
In order to overcome the defects of ESSC (Enhanced softsubspaceclustering), EWSC (Entropy Weighting subspaceclustering) and FWSC (Fuzzy Weighting subspaceclustering), a MOSSC (Multi-Objective Evolutionary-Based So...
详细信息
In order to overcome the defects of ESSC (Enhanced softsubspaceclustering), EWSC (Entropy Weighting subspaceclustering) and FWSC (Fuzzy Weighting subspaceclustering), a MOSSC (Multi-Objective Evolutionary-Based softsubspaceclustering) algorithm is proposed. Using multi-objective optimization technology, two objective functions of intra class and inter class in softsubspaceclustering method are optimized respectively. Using the method of two-partitioning of weighted subspaces, the optimal solution set of non-dominant Pareto is analyzed. The final clustering results are derived. Then, experiments were designed to apply the MOSSC algorithm on artificial datasets and real datasets. The results show that the MOSSC algorithm has better performance than the soft subspace clustering algorithm and the multi-objective clustering method. The partition effect of MOSSC algorithm is better than that of soft subspace clustering algorithm.
In recent years, the problem, which copes with high-dimensional data by the method of cluster analysis, has become a focus and difficulty in the field of artificial intelligence. Many conventional softsubspace cluste...
详细信息
ISBN:
(纸本)9781450365291
In recent years, the problem, which copes with high-dimensional data by the method of cluster analysis, has become a focus and difficulty in the field of artificial intelligence. Many conventional softsubspaceclustering techniques merge several criteria into a single objective to improve performance, however, the weighting parameters become important but difficult to set. A novel softsubspaceclustering with a multi-objective evolutionary approach (MOSSC) is proposed to this problem. First, two new objective function is constructed by minimizing the within-cluster compactness and maximizing the between-cluster separation based on the framework of soft subspace clustering algorithm. Based on this objective function, a new way of computing clusters' feature weights, centers and membership is then derived by using Lagrange multiplier method. The properties of this algorithm are investigated and the performance is evaluated experimentally using UCI datasets.
暂无评论