基于Fast Unfolding算法的情感词典扩展方法研究  

Research on Sentiment Dictionary Expansion Method Based on Fast Unfolding Algorithm

作　　者：陈力 黄树成[1] CHEN Li;HUANG Shucheng(School of Computer Science,Jiangsu University of Science and Technology,Zhenjiang 212003)

机构地区：[1]江苏科技大学计算机学院,镇江212003

出　　处：《计算机与数字工程》2023年第2期405-410,共6页Computer & Digital Engineering

基　　金：国家自然科学基金项目“基于鲁棒表现建模的目标跟踪方法研究”(编号:61772244)资助。

摘　　要：随着互联网技术的发展,越来越多的互联网产品出现了,极大地改变了人们的生活方式。越来越多的人选择在网络上发表自己的观点,分享自己的生活。这些行为带来了大量的数据,其中蕴含着丰富的商业和社会价值。对这些数据进行情感分析就是挖掘其背后隐藏价值的过程。论文在SO-PMI算法的基础上,提出将Fast Unfolding算法与PMI相结合,设计并实现FU-PMI算法,并使用准确率、召回率及F1值对算法的有效性进行评判。对比实验表明,论文提出的算法构建的情感词典相比于SO-PMI算法构建的情感词典效果更好,验证了FU-PMI算法的有效性。With the development of Internet technology,more and more Internet products have appeared,greatly changing people's lifestyles.People tend to choose to share their opinions and lives on the Internet.These behaviors bring a lot of data,which contains rich commercial and social value.The emotional research and analysis of these data is to excavate the information contained in datas.Based on the SO-PMI algorithm,this paper proposes to combine the Fast Unfolding algorithm with PMI to design and implement the FU-PMI algorithm and use the accuracy rate,recall rate and F1 value to judge the effectiveness of the algorithm.The experiment shows that the sentiment dictionary constructed by the algorithm proposed in this paper is better than the sentiment dictionary constructed by the SO-PMI algorithm,which verifies the effectiveness of the FU-PMI algorithm.

关 键 词：情感词典 情感分析 Fast Unfolding SO-PMI 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]