首页> 外文会议>IEEE International Conference on Software Engineering and Service Science >An improved algorithm of Chinese comments opinion mining based on adverbs
【24h】

An improved algorithm of Chinese comments opinion mining based on adverbs

机译:一种基于副词的中文评论意见挖掘的改进算法

获取原文

摘要

The existing opinion mining method based on adverbs about Chinese comments has lower precision rate and recall rate. For this problem, this paper redefined the narrow opinion word, and proposed an improved algorithm of Chinese comments opinion mining based on adverbs: In the extraction process of the feature words and opinion words based on adverbs increased conjunctions and negative words judgment; Then using the Extended Version of Synonymy Thesaurus of information retrieval lab in university of Harbin technology to merge synonyms. And obtaining the final feature words and opinion words. Experiments proved the method in this paper is better than the existing method. In opinion words extraction work the recall rate increased by 10.45%, the precision rate increased by 5.11%, the value of F increased by 0.08. In feature words extraction work the recall rate increased by 6.61%, the precision rate increased by 1.73%, the value of F increased by 0.04; In synonyms merger work the recall rate increased by 19.08%, the precision rate increased by 21.96%, the value of F increased by 0.09.
机译:现有的基于中文评论副词的意见挖掘方法准确率和召回率较低。针对这一问题,本文重新定义了狭义的意见词,并提出了一种改进的基于副词的中文意见舆论挖掘算法。然后利用哈尔滨工业大学信息检索实验室同义词同义词库的扩展版本来合并同义词。并获得最终特征词和意见词。实验证明,该方法优于现有方法。用词来说,提取工作的查全率提高了10.45%,准确率提高了5.11%,F值提高了0.08。在特征词提取工作中,召回率提高了6.61%,准确率提高了1.73%,F值提高了0.04;在同义词合并工作中,召回率增加了19.08%,准确率增加了21.96%,F值增加了0.09。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号