基于带特征染色体遗传算法的支持向量机特征选择和参数优化

鉴于支持向量机特征选择和参数优化对其分类准确率有重大的影响,将支持向量机渐近性能融入遗传算法并生成特征染色体,从而将遗传算法的搜索导向超参数空间中的最佳化误差直线.在此基础上,提出一种新的基十带特征染色体遗传算法的方法,同时进行支持向量机特征选择和参数优化.在与网格搜索、不带特征染色体遗传算法和其他方法的比较中,所提出的方法具有较高的准确率、更小的特征子集和更少的处理时间.

Feature selection and parameter optimization for SVM based on genetic algorithm with feature chromosomes

The classification accuracy of support vector machines(SVM) depends on feature selection and parameter
optimization of SVM strongly. The asymptotic behaviors of support vector machines are fused with genetic algorithm
and the feature chromosomes are generated, which directs the search of genetic algorithm to the straight line of optimal
generalization error in the superparameter space. On this basis, a new approach based on genetic algorithm with feature
chromosomes is proposed to simultaneously optimize the feature subset and the parameters for SVM. Compared with the
grid search, the genetic algorithm without feature chromosomes and other approaches, the proposed approach has higher
classification accuracy, smaller feature subset and fewer processing time.