Application of Chaotic Particle Swarm Optimization Algorithm in Chinese Documents Classification

In this paper, by using the ergodicity of chaos to improve the traditional particle swarm optimization algorithm, a chaos-PSO based hybrid optimization method is proposed. The core of document classification is constructing the classification model, the chaos PSO algorithm is utilized to extract classification rules so as to build the model rapidly. Michigan scheme is introduced to encode the rule, each particle can be viewed as a classification rule, the value of each particle is composed of document term weights. In the process of extracting classification rule with iterative optimization, the swarm is guided to chaotic search by altering the update strategy of particle´s location, which can make the algorithm get away from local optima and swell its capability to seek the global optimal solution, thereupon the categorization rules can be extracted accurately and effectively. Experiment results show that this method is feasible for Chinese document classification, it has good precision and high time efficiency.