Text Emotion Analysis based on Convolutional Neural Network and Naive Bayes

Abstract

In the era of Internet big data, how to preprocess non-standard text data and obtain effective classification features from it has an important impact on text emotion analysis(TEA). In the work of emotion analysis, text data preprocessing and emotion feature acquisition are the basis of emotion analysis task. For this reason, this paper proposes convolutional neural network(CNN) and NB algorithm technology to study TEA. Firstly, it introduces the deep learning based cyclic neural network model and NB algorithm technology, constructs a TEA model based on CNN and Naive Bayes(NB), and then designs the emotion analysis model according to the process of text evaluation object extraction; Finally, taking microblog as the research object, by comparing the experimental results of CRF model of machine learning, model based on cyclic neural network and model based on CNN and NBian model proposed in this paper, it is found that the emotion classification model designed in this paper combined with the model of Word2vec word vector has obtained the best experimental results, which proves the feasibility of this method.