特征选择是模式识别的关键问题之一,特征选择结果的好坏直接影响着分类器的分类精度和泛化性能.首先分析了特征选择方法的框架;然后从搜索策略和评价准则两个角度对特征选择方法进行了分析和总结;最后分析了对特征选择的影响因素,并指出了实际应用中需要解决的问题.
Feature selection is one of the key processes in pattern recognition.The accuracy and generalization capability of classifier are affected by the result of feature selection directly.Firstly,the framework of feature selection algorithm is analyzed.Then feature selection algorithm is classified and analyzed from two points which are searching strategy and evaluation criterion.Finally,the problem is given to solve real-world applications by analyzing infection factors in the feature selection technology.