Cluster ensemble algorithm using the Binary k-means and spectral clustering

Cluster ensemble has been shown to be an effective thought of improving the accuracy and stability of single clustering algorithms. It consists of generating a set of partition results from a same data set and combining them into a final one. In this paper, we develop a novel cluster ensemble method named Cluster Ensemble algorithm using the Binary k-means and Spectral Clustering (CEBKSC). By using the binary k-means algorithm and the spectral clustering method, the proposed method requires low computational complexity and is therefore very suitable for large text data sets. It works by firstly using the binary k-means algorithm to create a set of partition results and then integrating these results by using the spectral clustering. In addition, we introduce a matrix transformation technique to lower the computational cost of the spectral clustering. Experiments show that the proposed method has better clustering quality and is faster than several other cluster ensemble methods. 1553-9105/