摘要
随着微博技术的不断发展,越来越多的用户喜欢在微博上发表自己对某件事或某个问题的观点看法。找出一个可以有效的判断评论对某事件所持态度的方法也日益引起人们的关注,观点信息抽取也成为一个重要的研究课题。文章提出的是一种自动识别评论态度的方法,主要是用微博文本中表情动画、动词以及句型等特征信息来构造特征集,并根据这个特征集生成特征向量,利用SVM分类器生成测试模型。实验结果表明,该方法取得了较好的系统性能。
With the continuous development of microblog technology, more and more people would like to give their views on some hot topics in their microblog, and thus how to get the information about the attitude of comments to hot topics receives much attention from the people. Opinion extraction now becomes one of the important research subjects. This paper proposes a systemic project about distinguishing the viewpoints of the comments. In this work, the features like emoticons in microblogs, verb and sentence patterns etc should be picked up to construct a feature set, then feature vectors are produced, and test model is made by using SVM classifier. The experimental results show that this method is of high system performance.
出处
《信息安全与通信保密》
2013年第1期49-50,53,共3页
Information Security and Communications Privacy
