Affective Analysis of Chinese Sentences Based on Word2vec and SVC

Abstract

This Paper was based on word2vec, Using Chinese Encyclopedia Corpus of Wikipedia as the training set to generate Chinese word vector and Chinese sentence vector, and using SVC (support vector classification) to classify the text of 16542 comments in a hotel industry, that is, to realize the affective analysis of Chinese sentences. The results show that the sentence vectors generated by the voting SVC model are better than those generated by the mean SVC model.