摘要：伴随着移动的网络终端发展，自新浪微博推出以来，中文微博使用者量巨增，微博数据量也呈现出指数级的增长。大量的微博数据中隐藏了巨量的内容情感，我们对这些数据进行情感分析就可以得到有用的支持信息去决策。针对微博文本文字进行情感分析，这样有利于情感的发现、监控以及进行信息预测和商品销售、产品改进等。中文微博与普通文本文字的区别，有其特殊性，相应的情感分析方法也区别于普通的文本文字情感分析。目前存在微博情感的分析方法有优点也有缺点，为了提高对中文微博的情感分析准确度，我们主要进行了以下两方面研究。其中情感词的集合是情感词典，是表达情感的重要特质，构建完善的情感词典对微博查找是至关重要的，提出了基于表情符号的新词的识别方法。综合程度、情感词和否定副词之间的相互影响作用，总结归纳了七类关于副词和情感词之间修饰方式的语义规律，创新的添加了后置否定词的语义规则。 42519

毕业论文关键词：网络终端；微博；情感；语义规则

Chinese microblog emotion Analysis

Abstract: In recent years, with the development of mobile Internet terminal, since microblog was push out,Chinese microblog user growth and has also shown an exponential growth in the volume of data. Lots of microblog data hidden in the large amount of emotion,sentiment analysis of these data can be useful support information to the decision. Aim at microblog text words to sentiment analysis is good at emotion of discovery, monitoring, and forecast the information and merchandise sales,product improvement etc。Differences between Chinese microblog and normal text in Chinese, has its particularity, the corresponding emotion analysis methods are also different from ordinary text sentiment analysis. There are microblog emotion analysis method has both advantages and disadvantages, we conducted two studies in order to improve the accuracy of the Chinese microblog sentiment analysis. Microblog emotion expression elements were analyzed by totals and give it weight, including emotional words,degree, emoticons, negative adverbs, adverb,as well as the related sentence patterns, patterns. Emotion dictionary is a collection of the words of which is an important trait of emotional expression, build a comprehensive dictionary of feelings on microblog find is crucial to the proposed new relationship of word recognition method based on emoticons. Summarizing the seven categories on adverb and finishing between emotional Word semantic rules, innovative post negative semantic rules have been added.

Key words: Network Terminal; microblog; emotion ; Semantic rules

目    录

摘  要 1

引言 2

1.研究的背景和意义 2

2.课题研究现状 3

3.情感分析 4

4相关理论和运用技术 4

4.1对于停用词的过滤 4

4.2对于文字特征的抽取 5

4.3文本文字预处理及表示 5

4.4分词和词类性质的标注 5

4.5贝叶斯分类算法[10] 5

5 中文微博情感分析方法 6

5.1基于情感词典的分析方法 6

    5.1.1情感倾向判断的基本规则 6

    5.1.2 表情符号词典 7