论文部分内容阅读
文章以中文微博为研究对象,结合心理学和自然语言处理,将微博情绪划分为乐、怒、哀、恶、惧五大类。然后在类别划分的基础上,使用情感特征、句式特征、句间特征来表示微博情绪,并借助于SVM模型形成了微博情绪分类模型。最后借助NLP&CC 2013的公开评测数据对提出的模型进行了验证,实验结果表明本文所提的方法是有效的。
The article takes Chinese microblogging as the research object and combines the psychology and natural language processing to classify the emotion of the microblog into the five categories of music, anger, sadness, evil and fear. Then, based on the classification, we use the emotion features, the sentence features and the inter-sentence features to represent the micro-blog emotion, and form the micro-blog emotion classification model by means of SVM model. At last, the proposed model is verified by the public evaluation data of NLP & CC 2013, and the experimental results show that the proposed method is effective.