Particle swarm optimization-based text characteristic selection method

The invention discloses a particle swarm optimization-based text characteristic selection method. The method aims to solve the problems of high dimension and sparsity of a text eigenvector due to adoption of a spatial vector model for representing a text. A local search strategy is embedded into a particle swarm optimization algorithm to select out unrelated and significant characteristic subsets;and the particle swarm algorithm is guided to select different characteristics in a search process by considering relevant information of a particle swarm, thereby selecting out the characteristics more favorable for classification accuracy from original characteristics. The characteristic subsets most favorable for text representation can be selected out from a huge text word set, thereby layinga good foundation for text classification and processing. 本发明公开了种基于粒子群优化的文本特征选择方法，该方法是为了解决采用空间向量模型表示文本出现文本特征向量高维且稀疏的问题，本发明将局部搜索策略嵌入到粒子群优化算法中选择出不相关和显著的特征子集，通过考虑粒子群的相关信息来指导粒子群算法在搜索过程中选择不同的特征，从而从原始特征中选择出更加有利于分类准确率的特征。本发明能够从庞大文本词