论文部分内容阅读
为了在特征选择中获得具有较高分类准确率的特征子集,提出了一种基于支持向量机递归特征消除法(SVM-RFE)和二进制粒子群算法(BPSO)的特征选择方法.该方法首先利用SVM-RFE快速去掉部分无关特征,初步缩减数据维数,然后以粒子群算法继续搜索最优子集,并将SVM-RFE算法得到的优良子作为粒子群算法的部分初始种群,使后续粒子群算法有一个较好的搜索起点.SVM-RFE既减少了粒子的搜索空间,又为其提供了先验知识,从而提高算法的搜索效率和识别精度.实验结果表明,该方法可以在分类准确率更高或相等的情况下得到维数更少的子集.
In order to obtain feature subsets with higher classification accuracy in feature selection, a feature selection method based on support vector machine recursive feature elimination (SVM-RFE) and binary particle swarm optimization (BPSO) is proposed. Firstly, The SVM-RFE is used to quickly remove some irrelevant features and reduce the data dimension. Then, the optimal subsets are continuously searched by the PSO algorithm, and the good children obtained by the SVM-RFE algorithm are used as part of the initial population of the PSO. Swarm algorithm has a good search starting point.SVM-RFE not only reduces the particle search space, but also provides a priori knowledge to improve the search efficiency and recognition accuracy of the algorithm.The experimental results show that this method can be used in the classification accuracy A higher or equal rate results in a smaller subset of dimensions.