基于聚类和二进制PSO的特征选择

A Novel Algorithm Based on K-Means Clustering and Binary Particle Swarm Optimization

特征选择是模式识别及数据挖掘等领域的重要问题之一。特征选择不但可以提高分类精度和效率,也可以找出富含信息的特征子集。针对此问题,在分析了常用的一些特征选择算法之后,文中提出一种基于聚类和二进制PSO算法的特征选择方法,首先基于特征之间的相关性聚类来进行特征分组及筛选,然后针对经过筛选而精简的特征子集采用二进制粒子群算法进行随机搜索。实验结果表明,该算法可有效地找出具有较好的线性可分离性的特征子集,具有特征精简幅度较大、运行效率较高等优点。

Feature selection is one of the important problems in the pattern recognition and data mining areas.For high dimensional data feature selection not only can improve the accuracy and efficiency of classification,but also can discover informative feature subset.The new feature selection method combining k-means and PSO was proposed in this paper,which first filters feature by k-means,and realized the near optimal feature subset search on the compact feature subset by PSO algorithm.The experiments show that the proposed algorithm can get a good compact feature subset and run more efficiently.