摘要
在对中文文本特点进行分析的基础上,提出了一种新型的情感倾向计算模型。该模型采用了改进的逐点分析方法SO-PMI(Semantic Orientation from Pointwise Mutual Information),基于词语对类别的倾向性进行分类。实验表明,该模型在汉语常用词中的效果较好,判别准确率可达80%以上,具有一定的实用价值。
On the basis of the analysis of characteristics of the Chinese text,the thesis introduces a novel emotional tendency computation model.The model adopts an improved SO-PMI to classify emotional tendencies based on words.Experiments show the model works accurately with frequently used Chinese words with a hit rate of above 80 percent.Therefore it is of practical value.
出处
《计算机应用与软件》
CSCD
2011年第6期271-272,285,共3页
Computer Applications and Software
关键词
文本倾向性分类
中文信息处理
态度分类
语义倾向
Text tendency classification Chinese information processing Sentiment classification Semantic tendency
