基于SVM的网络文本情感分类系统的研究与设计  

Research and design of network text sentiment classification system based on SVM

作　　者：樊康新[1] 

机构地区：[1]南通大学计算机科学与技术学院,江苏南通226019

出　　处：《计算机时代》2015年第12期34-37,共4页Computer Era

基　　金：南通大学自然科学基金资助项目(11Z075)

摘　　要：文本情感分类是指通过分析文本中的立场、观点、态度、情绪等主观信息,对文本的情感倾向做出类别判断。设计了一个基于SVM的网络文本分类系统;描述了该系统的结构和工作流程;对该系统实现的关键技术包括情感词典的构建、情感特征的选择、情感值的计算、网络文本情感分类器的构造等进行了深入的分析和研究。实验结果表明,该系统对网络文本情感倾向分类已达到较为理想的效果。Text sentiment classification automatically classifies a text as expressing positive or negative sentiment through analyzing subjective information in the text such as standpoint, viewpoint, attitude, mood, and so on. A network text classification system based on SVM is designed in this paper. The system structure and work flow are described, and the key techniques are particularly discussed, that are how to construct the sentiment lexicon, how to select sentiment feature, how to calculate sentiment value, and how to construct the network text sentiment classifier, etc. The experiment results show that the system has the ideal effect on the classification of the sentiment orientation of Internet text.

关 键 词：情感分类 网络文本 SVM 情感分析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]