基于三支决策的中文微博观点句识别研究

发布时间：2018-02-26 10:41

本文关键词： 三支决策中文微博观点句　出处：《山东大学学报(理学版)》2014年08期 　论文类型：期刊论文

【摘要】：微博观点句识别是对微博进行观点挖掘和舆情分析的基础,因此观点句识别的准确率对后续研究工作至关重要。提出了一种基于三支决策的中文微博观点句识别方法,采用支持向量机分类器,sigmoid函数计算某条微博属于观点句的概率,并将基于三支决策的中文微博观点句识别方法与传统支持向量机方法进行对比实验,实验结果表明,基于三支决策的中文微博观点句识别方法取得了很好的识别效果。
[Abstract]:Weibo's opinion sentence recognition is the basis of the opinion mining and public opinion analysis, so the accuracy of opinion sentence recognition is very important to the follow-up research. The support vector machine classifier sigmoid function is used to calculate the probability of Weibo's opinion sentence, and a comparative experiment is made between the traditional support vector machine (SVM) and the Chinese Weibo's opinion sentence recognition method based on three-line decision. The experimental results show that, The method of Chinese Weibo's opinion sentence recognition based on three-branch decision has achieved good recognition effect.
【作者单位】：湖南工业大学计算机与通信学院;
【基金】：国家自然科学基金资助项目(61170102) 国家社科基金资助项目(12BYY045)
【分类号】：TP393.092

【参考文献】