加权维吾尔语句子倾向性分析

准确可靠的文本倾向性分析是网络舆情分析与网络内容安全的前提.本文提出了利用中文极性情感词典HowNet、NTUSD以及大连理工大学发布的褒贬情感词词典进行并交运算，选择并翻译为维吾尔语词汇，借助于维吾尔语同义近义词词典，扩展构建了维吾尔语极性情感词典；然后分析总结了否定词、程度副词以及句中的转折连词等情感修饰成分对维吾尔语句子情感极性的影响，并量化为情感词权值；最后设计了基于维吾尔语极性情感词和权值相结合的加权句子情感极性判定算法.利用自建语料库进行测试，并与汉语倾向性判定实验结果比较，证明了本算法进行维吾尔语句子褒贬情感性分析基本是有效地.

Analysis of the Sentence Tendency in Uighur Language

The precondition of network public opinion analysis and network content security is based on the accurate and reliable text tendency analysis. In this paper the Chinese emotion dictionary which included HowNet, NTUSD and emotion word dictionary released by Dalian university is implemented intersection operation and union Operation. By selecting, translating for Uighur vocabulary, with the synonymous dictionary, the Uighur emotional polarity dictionary is constructed. At the same time the impact is analyzed and summarized about the Emotional Modifier such as the effect of negative word, degree adverbs and the Sentence''s adversative conjunction etc. Then the effect degree is converted into emotional weight value. In the end the weighted algorithm based on the emotion words and the modifier words is designed to decide a sentence emotional polarity. The experiment result proves that this algorithm is basic validity compared to Chinese emotional tendentiousness algorithm by testing on self-built corpus.