A Text Clustering Method Based on Two-Dimensional OTSU and PSO Algorithm

Fast and high-quality text clustering algorithm is an important and challenging problem in effectively navigating. Such as the high-dimensional sparse text data, poor efficiency of unsupervised feature selection, and defects existing in classical clustering methods and so on. In this paper, an effective and unsupervised text clustering method (OK-PSO) is proposed. First, k-means is used to calculate the distance from each term to the cluster centers, and then the two-dimensional Otsu algorithm is included to evaluate the optimization of clustering distance threshold. The process of 2D Otsu is taken by PSO algorithm. Finally, several experiments are taken based on OK-PSO and some other methods. The experimental results illustrate the efficiency of OK-PSO method proposed in this paper.