0942152116-彭灿《嵌入式系统及应用》大作业模版.doc

《嵌入式系统及应用》 （） 题目： 姓 名： 彭灿 学 号： 0942152116 专 业： 院 系： 工程学院 指导老师： 完成时间： 2013年月日 基于意见挖掘的中文词汇语义倾向性分析 摘 要 近年来互联网的发展，网民言论将对政府在内的各类社会组织发挥前所未有的影响力。对这类网络进行分析政府了解的。如何处理这些信息从中得到他们想要的内容，在近年来得到了很多的关注和研究。其中情感分析就是随之兴起的一个研究领域，可以应用到信息过滤、产品推荐和用户兴趣发掘等方面。 本文主要针对热点新闻事件中的评论数据进行收集，然后利用中科院分词软件TCTCLAS对收集到的评论数据进行分词，将它们看成一段段的短文本，抽取其中的关键词（一般以形容词、副词居多），与事先准备好的情感语料库进行对比，提出语义倾向相似度的概念，通过对PMI算法和HowNet算法计算词语的倾向性的优缺点进行分析，提出一种改进的基于意见挖掘的中文词汇语义倾向性分析方法。从而为相关部门针对这些问题实施新的政策提供更好的决策支持关键词：情感倾向性；文本倾向性分析；情感分析Based On Analysis Of Chinese Lexical Semantic Tendentiousness Opinion Mining ABSTRACT With the rapid development of the network, Internet has become an important place where people could express their opinions. Internet commenting also became a vital method where the government and enterprises get people’s opinions and customers’ need tendency which will make great and immeasurable impacts on the government and all kinds of social organizations. The government can get to know netizens’ view of public affairs, and enterprises can adjust and promote their products quality and after services through the analysis of comments posted online. In recent years, more and more researches have focused on the topic about how to filer and select information they need. And emotion analysis is one of those researches which could be applied in information filtering, products recommending, user interest developing and many other aspects. This paper is a research concentrating on the analyzing of the tendency of Chinese words based on comment discovering. The paper is mainly about: 1. Basic concepts about comment discovering, and a deep introduction about related researches both from China and foreign countries. 2. An introduction to two basic calculating methods in present calculations of words tendency, and the analysis of advantages and disadvantages of