基于自适应t分布变异的粒子群特征选择方法

Feature selection algorithm using PSO with adaptive mutation-based t distribution

在分析粒子群优化(particle swarm optimization,PSO)的基础上,提出了一种基于自适应t分布变异的简化粒子群特征选择方法。针对PSO容易陷入局部收敛的缺陷,通过对群体极值进行自适应t分布变异,使其跳出局部收敛。为了解决随机选择初始群体可能会延长搜索时间这一问题,将互信息引入到算法中。通过计算特征与类别的相关性来确定每个特征的入选概率,根据概率值生成一个近似最优粒子,使粒子群一开始就沿着比较合理的方向搜索,从而缩短进化时间。最后,以支持向量机(support vector machine,SVM)为分类器,通过仿真实验验证了算法的可行性和有效性。

By analyzing particle swarm optimization （PSO）, a feature selection algorithm is proposed which is on the ground of the simple PSO with adaptive mutation-based t distribution. As PSO relapses into local ex- tremum easily, the adaptive mutatiowbased t distribution is used to break it away from local extremum effec- tively. Meanwhile, the mutual information is introduced for fear that initializing particles selected randomly may ex- tend the searching time. The choosing probability of each feature is obtained by computing the dependence between the feature and class, by means of which an approximate optimum particle is generated. Therefore, the particle swarm can search along the reasonable direction and the evolutionary time is saved. Experiment results indicate the feasibility and validity of the algorithm with a support vector machine （SVM） as the classifier.