When using spectral clustering algorithm to perform clustering, there are some shortcomings, such as slow convergence rate, and the clustering result is easily affected by the initial center. In order to improve this problem, this paper proposes an improved spectral clustering algorithm based on cell-like P system, called SCBK-CP algorithm. Its main idea is to use the bisecting k-means algorithm instead of k-means algorithm and construct a cell-like P system as the framework of the bisecting k-means algorithm to improve the spectral clustering algorithm. The maximum parallelism of the P system improves the efficiency of the bisecting k-means algorithm. The algorithm proposed in this paper improves the clustering effect of spectral clustering, and also provides a new idea for the application of membrane computing. The SCBK-CP algorithm uses three UCI datasets and an artificial dataset for experiments and further comparison with traditional spectral clustering algorithms. Experimental results verify the advantages of the SCBK-CP algorithm.
展开▼
