基于Word2Vec的维吾尔语情感词典扩展研究

The Uighur Emotional Dictionary Extension Based on Word2vec

情感词典是情感识别的基础资源,在前期构建维吾尔语情感词典的基础上,基于Word2Vec开展情感词的自动扩展研究。首先,根据维吾尔语情感词的语言特征从搜集的维吾尔语网络语料中提取候选情感词;其次,利用Word2Vec模型训练维文语料,以获取词语的向量表示;最后,在此基础上,利用基础情感词典资源和表征语义相似度的向量余弦值算法来判定候选情感词的情感指向。

Emotion dictionary is the basic resource of emotion recognition. Based on the construction of Uyghur emotion dictionary in the early stage, this paper carries out the research on the automatic expansion of emotion words based on Word2Vec. Firstly, according to the language characteristics of the emotion words in Uyghur language to extract candidate emotional words from the collected Uyghur language network corpus. Secondly, using Word2Vec weaving model training corpus obtain the word vector said. Finally, on this basis, an algorithm using the basic emotion dictionary resources and the vector cosine value representing the semantic similarity is designed to determine the emotion direction of the candidate emotion words.