Chinese Word Segmentation based on the Improved Particle Swarm Optimization Neural Networks

The Chinese word segmentation based on the improved Particle Swarm Optimization (PSO) neural networks is discussed in this paper. Firstly, a solution is obtained by searching globally using FPSO (Fuzzy cluster Particle Swarm Optimization) algorithm, which has strong parallel searching ability, encoding real number, and optimizing the training weights, thresholds, and structure of neural networks. Then based on the optimization results obtained from FPSO algorithm, the optimization solution is continuously searched by the following BP algorithm, which has strong local searching ability, until it is discovered finally. Simulation results show that the method proposed in this paper greatly increases both the efficiency and the accuracy of Chinese word segmentation.