| 文本分类中一种特征选择方法的改进 |
| |
| 引用本文： | 胡改蝶,马建芬.文本分类中一种特征选择方法的改进[J].计算机与现代化,2011(5):20-21,24. |
| |
| 作者姓名： | 胡改蝶 马建芬 |
| |
| 作者单位： | 太原理工大学计算机科学与技术学院,山西,太原,030024 |
| |
| 摘 要： | 特征选择在文本分类中是非常必要的,这是由于它可以使分类更加有效与准确。本文根据特征选择方法χ2统计方法的不足,对χ2统计进行改进,并在支持向量机分类算法上进行实验。实验结果表明改进的方法可以提高分类的准确度。
|
| 关 键 词： | 特征选择 文本分类 χ2统计 支持向量机 |
Improvement of Feature Selection Method in Text Classification |
| |
| HU Gai-die,MA Jian-fen.Improvement of Feature Selection Method in Text Classification[J].Computer and Modernization,2011(5):20-21,24. |
| |
| Authors: | HU Gai-die MA Jian-fen |
| |
| Affiliation: | HU Gai-die,MA Jian-fen(College of Computer Science and Technology,Taiyuan University of Technology,Taiyuan 030024,China) |
| |
| Abstract: | In text classification,feature selection is essential to make classification efficient and more accurate.This paper depends on deficiency of χ2 statistics and presents an improved χ2 statistics method.Experiments based Support Vector Machine show that the improved method can improve the classification accuracy. |
| |
| Keywords: | feature selection text classification χ2statistics support vector machine |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
