Sentiment Analysis by Exploring Large Scale Web-based Chinese Short Text

Abstract

Sentiment analysis has been a popular research topic these years. In this paper, we will pay attention to the sentimental state of college students and make corresponding sentiment analysis. In this paper, we purpose two short text sentiment analysis models based on student forum data, named hard mapping sentiment analysis model and soft mapping sentiment analysis model respectively. Hard mapping sentiment analysis obtains the polarity of each short text by Polarity Value Table of Chinese Sentimental Words. On this basis, soft mapping sentiment analysis model trains the word similarity model of Chinese words with a large amount of data from Chinese Wikipedia using word2vec and Gensim. It diminishes the judgment errors induced by the inadequate vocabulary size of the sentiment dictionary in the Hard mapping sentiment analysis. Our result shows the improvement in short text sentiment analysis of the model by using the soft mapping sentiment analysis model. The research has significant implications in the psychological guidance of college students.