类别区分词与情感词典相结合的维吾尔文句子情感分类

发布时间：2018-04-13 21:00

本文选题：类别区分词 + 情感词典　；参考：《清华大学学报(自然科学版)》2017年02期

【摘要】：该文在类别区分词特征选择方法的基础上,针对维吾尔文中的生气、高兴、难过、惊讶等句子的情感类别提出了类别区分词与情感词典相结合的方法,进行了句子情感分类研究。结合维吾尔语文本句子中的情感表达特点,利用类别区分词特征选择方法,提取了最有类别区分能力的特征词,并进行了情感分类。通过人工抽取方法收集了维吾尔文句子中能表达情感的关键词,并建立了一个基础情感词典。将该词典与类别区分词结合在一起作为特征,对维吾尔文句子的情感类型有效地进行了分类。实验结果表明类别区分词与情感词典相结合方法的分类效率优于只用类别区分词特征选择方法。
[Abstract]:Based on the feature selection method of category distinguishing words, this paper proposes a method of combining category distinguishing words with emotion dictionaries for the emotional categories of Uygur sentences, such as anger, joy, sadness, surprise, etc.A study of sentence affective classification is carried out.Based on the characteristics of emotion expression in Uygur Chinese sentences, the feature selection method of category distinguishing words is used to extract the feature words with the most ability of category differentiation, and the emotion classification is carried out.The key words in Uygur sentences are collected by artificial extraction, and a basic emotion dictionary is established.By combining the dictionary with category distinguishing words, the emotion types of Uygur sentences are classified effectively.The experimental results show that the classification efficiency of the combination of category distinguishing words and affective dictionaries is better than that of only category distinguishing words feature selection methods.
【作者单位】：新疆大学软件学院;新疆大学信息科学与工程学院;
【基金】：国家“青年千人计划”项目国家自然科学基金项目(61163033;61065005) 教育部新世纪优秀人才支持计划资助项目(NCET-10-0969) 新疆维吾尔自治区高新技术发展研究计划项目(201312103)
【分类号】：TP391.1

【相似文献】