分类问题中过滤式选择建模属性的方法及应用

Method and Application of Selecting Modeling Attributes by Filtering in Classification Problem

属性是描述数据特征的基本单元.首先分析属性选择的意义,继而阐述了过滤式属性选择的原理,结合垃圾短信样本数据进行实证分析,结果表明,属性筛选对于建模至关重要,经过属性筛选,模型准确率有了大幅提升.

Attributes are the basic unit for describing data features.This paper first analyzes the meaning of attribute selection,and then expounds the principle of attribute selection by filtering,and finally carries out an empirical analysis in combination with the spam sample data.The results show that attribute selection is critical for modeling.