Binary Information Press
Text clustering is one of the core technologies of text mining and information retrieval. But, the performance of traditional clustering algorithm may not be satisfying in clustering text data due to the high dimensionality and sparseness of text data. This paper presents a partitioned clustering approach based on the statistic of the text dataset and PSO (Particle Swarm Optimization) to handle the problem of text clustering. The dense regions of dataset were selected out as cluster centroid sets gradually by calculating the similarity information between the already partitioned and left sets.