| 结合AB-SMOTE和C-SVM的中文倾向性句子识别 |
| |
| 引用本文： | 陈振伟,廖祥文.结合AB-SMOTE和C-SVM的中文倾向性句子识别[J].福州大学学报(自然科学版),2012,40(3):310-315. |
| |
| 作者姓名： | 陈振伟 廖祥文 |
| |
| 作者单位： | 福州大学数学与计算机科学学院 |
| |
| 基金项目： | 福建省自然科学基金资助项目(2010J05133);福建省科技创新平台计划资助项目(2009J1007);福州大学科技发展基金资助项目(2010-XQ-22) |
| |
| 摘 要： | 提出一种结合AB-SMOTE和C-SVM的中文倾向性句子识别算法.该算法先利用AB-SMOTE方法合成新样本,降低不平衡程度的同时也使数据具有更好的代表性,再对不同类别赋予不同的惩罚系数形成代价敏感的C-SVM分类器,充分结合了数据层和学习算法层方法的优点.实验结果表明,对酒店、笔记本电脑和书籍3个不平衡语料处理时,本算法能较有效解决不平衡问题,提高倾向性句子的识别精度.
|
| 关 键 词： | 中文 倾向性句子 识别 算法 AB-SMOTE C-SVM |
Identification of Chinese opinion sentence combined AB-SMOTE with C-SVM |
| |
| CHEN Zhen-wei,LIAO Xiang-wen.Identification of Chinese opinion sentence combined AB-SMOTE with C-SVM[J].Journal of Fuzhou University(Natural Science Edition),2012,40(3):310-315. |
| |
| Authors: | CHEN Zhen-wei LIAO Xiang-wen |
| |
| Institution: | (College of Mathematics and Computer Science,Fuzhou University,Fuzhou,Fujian 350108,China) |
| |
| Abstract: | This paper proposes a new method that identification of opinion sentences combined AB SMOTE and C-SVM.Firstly,use AB-SMOTE to synthesize new samples of small class,it not only reduce imbalance degree but also make the data has better representativeness.Then,give different cost coefficient to different class and get C-SVM.The method has sufficiently combined the advantage in data-level and algorithm-level.The experiments show that is relatively efficient to deal with the imbalance dataset problem and can get higher accuracy. |
| |
| Keywords: | Chinese opinion sentence identification algorithm AB-SMOTE C-SVM |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《福州大学学报(自然科学版)》浏览原始摘要信息 |
| 点击此处可从《福州大学学报(自然科学版)》下载免费的PDF全文 |
