Network Public Comments Sentiment Analysis Based on Multilayer Convolutional Neural Network
Ming Jiang, Liqiang Jin, Feiwei Qin, Min Zhang, Ziyang Li · 2016
With the rapid development of the Internet, network review information shows explosive growth. How to make an accurate analysis of these web comments has become an important issue in the research of network public opinion. In this work, a traditional dictionary based algorithm is adopted to analyze text sentiment of these comments. A deep learning method is proposed to analyze the network public sentiment of these comments due to the inherent defects of the traditional method itself. Firstly a convolutional neural network(CNN) model is constructed as sentiment classifier, Secondly the words are translated into a space vector with the aid of Word2Vec. Hence the sentences are converted into a two-dimensional matrix as input of the classifier. Finally the deep learning method and the dictionary method are compared and analyzed. Experimental results show that the proposed deep learning approach is more accurate than the dictionary approach in analyzing the text sentiment of network public comments.