基于条件随机场模型的在线评论情感极性强度分析  

Analysis of Emotional Polarity Intensity of Online Comments Based on Conditional Random Field Model

作　　者：刘伟 李秀霞 张伟志 LIU Wei;LI Xiuxia;ZHANG Weizhi

机构地区：[1]曲阜师范大学传媒学院,山东日照276826

出　　处：《图书情报导刊》2021年第12期25-30,共6页Journal of Library and Information Science

摘　　要：为了辨别在线评论的情感倾向,了解大众舆情对某一事件或产品的看法,采用层叠CRFs模型,把在线评论语句进行主客观分类、极性分类以及评论语句的褒贬强度分类,充分利用极性分类和褒贬强度分类之间的层次关系来改善情感分类的冗余关系,以期借助改进的条件随机场模型实现文本情感倾向及强度的分析,实现篇章级文本集的情感分析。In order to identify the emotional tendency of online comments and understand the public opinion on a certain event or product,this paper uses the hierarchical CRFs model to classify online comment statements into subjective and objective categories,polarity categories and commendatory or negative intensity categories,makes full use of the hierarchical relationship between polarity categories and commendatory or negative intensity categories to improve the redundant relationship of sentiment classification,and with the help of the improved conditional random field model,analyzes the sentiment tendency and intensity of text,and to achieves the sentiment analysis of text sets at the discourse level.

关 键 词：情感分析 数据挖掘 在线评论 层叠CRFs模型 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]