Research on the Key Technology of Chinese Text Sentiment Analysis

In the era of big data, text sentiment analysis is of great significance to the analysis of public opinion. In general, there are two broad approaches on sentiment analysis, lexicon-based and machine learning-based method. In fact, sentiment analysis belongs to the classification technique as well. Therefore, this paper also studied the method based on deep learning. This paper implemented three approaches and compared the performances of different classification effects. The concept of Word2vec was also introduced to the machine learning-based method. The word vectorization method was used to extract the corpus features and reduce the dimension through the Principal Component Analysis (PCA)algorithm. The fully connected neural network was selected in deep-learning-based method. This paper used keras library to build neural network framework. By comparing the three methods, it was concluded that the machine learning method was the best. The correct rate was 85.60%.