Soft subspace clustering using QPSOSC algorithm

Yangyang Li, Xiaoxu Liang, Yujing Lu, Licheng Jiao

Abstract: Due to the sparseness and redundancy of high-dimensional data, clusters may exist only in some low-dimensional subspaces. And the similar method of measurement of the classical methods, such as K-means, is no longer applicable. So it is difficult to predict the result of high-dimensional data with traditional clustering. In this paper, the traditional cluster algorithm with its instability, easily trapping in local optimum has been modified and improved a lot. The Quantum-behaved Particle Swarm Optimization (QPSO) algorithm is introduced and combined with subspace clustering algorithm to optimize the weight matrix. This paper proposes a soft subspace clustering based on QPSO (QPSOSC) algorithm which effectively improves the diversity and stability of weight matrix. Experiment results of both synthetic and real data shown that QPSOSC outperformed several well-known algorithms, such as ESSC, FWKM, EWKM and LAC in almost all experiments. From the experiment results, it is obvious that QPSOSC not only improves the stability of the algorithm and the accuracy of the results effectively, but also reduces the possibility of falling into the local optimum.