| 一种基于遗传操作和类内类间距离判据理论的特征选择方法 |
| |
| 引用本文： | 孙雷,王新.一种基于遗传操作和类内类间距离判据理论的特征选择方法[J].计算机工程与应用,2004,40(21):178-181. |
| |
| 作者姓名： | 孙雷 王新 |
| |
| 作者单位： | 石油大学计算机科学与技术系,北京,102249 |
| |
| 摘 要： | 在分类挖掘的预处理过程中,重要的是如何选择最有效的特征子集,以便于分类学习与预测。该文提出一种基于遗传操作和类内类间距离判据理论的特征选择方法,该方法能够有效地缩减特征空间的维数,提高分类挖掘的效率。
|
| 关 键 词： | 特征选择 数据缩减 分类 |
| 文章编号： | 1002-8331-(2004)21-0178-04 |
A Feature Selection Method Based on Genetic Algorithm Options and Euclidean Distance among All Instances of Different Class |
| |
| Sun Lei Wang Xin.A Feature Selection Method Based on Genetic Algorithm Options and Euclidean Distance among All Instances of Different Class[J].Computer Engineering and Applications,2004,40(21):178-181. |
| |
| Authors: | Sun Lei Wang Xin |
| |
| Abstract: | During the preprocessing of classification,it is important to select feature subset effectively,in order to classification learning and predicting.In this paper a feature selection method is introduced,which is based on genetic algorithm options and Euclidean distance among all instances of different class.It can reduce effectively dimensions of feature space and enhance the efficiency of classification. |
| |
| Keywords: | feature selection data reduce classification |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
