特征选择方法与算法的研究  

Research on Feature Selection Methods and Algorithms

作　　者：李敏[1] 卡米力.木依丁 

机构地区：[1]新疆大学信息与科学工程学院,新疆乌鲁木齐830046

出　　处：《计算机技术与发展》2013年第12期16-21,共6页Computer Technology and Development

基　　金：国家自然科学基金资助项目(61065001);新疆少数民族科技人才特殊培养计划(201023116)

摘　　要：特征选择的主要思想是通过去除一些包含少量或不相关的信息的特征去选择特征子集。特征选择方法可分为三大类:一是过滤式,二是封装式,三是嵌入式。鉴于目前存在大量的特征选择算法,为了能够适当地决定在特定的情况下使用哪种算法,需要提出可以依赖或判定的标准。文中的主要工作就是综述一些基本特征选择算法,根据文献中已有的理论和实验结果对特征选择方法和算法进行比较分类,然后提出一种可以依赖或判定的标准。The main idea of feature selection is to choose a subset of input features by eliminating features with little or no predictive in- formation. Feature selection methods can be decomposed into three broad classes ： one is filter methods, another one is wrapper methods and the third one is embedded methods. In view of the substantial number of existing ＇feature selection algorithms, count on criteria that enable to adequately decide which algorithm to use in certain situations need arises to. This work reviews several fundamental algorithms found in the literature,proposed a criteria which guide researchers to make a decision to use proper algorithms by presenting an empirical comparison of feature selection methods and its algorithms.

关 键 词：特征选择方法 特征选择算法 过滤式 封装式 嵌入式 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]