Sentiment Analysis Algorithm Based on BERT and Convolutional Neural Network

Rui Man, Ke Lin

20212021 IEEE Asia-Pacific Conference on Image Processing, Electronics and Computers (IPEC)44 citationsDOI

Abstract

The explosion of Internet information makes timely analysis and monitoring of Internet public opinion more and more important, and sentiment analysis of public opinion events is even more important. The traditional word2vec model cannot fully express the information contained in words. It is proposed to use the BERT model as the article feature extraction model; and use the deep convolutional neural network to extract the local information of the article, and then connect the fully connected network to classify the article, so as to achieve sentiment analysis purpose. Experimental results on public data sets show that sentiment analysis algorithms based on BERT and convolutional neural networks are better than traditional sentiment analysis algorithms.

Topics & Concepts

Sentiment analysisComputer scienceWord2vecConvolutional neural networkArtificial intelligenceThe InternetDeep learningPublic opinionArtificial neural networkFeature extractionFeature (linguistics)Machine learningData miningWorld Wide WebPoliticsLawEmbeddingPhilosophyLinguisticsPolitical scienceSentiment Analysis and Opinion MiningAdvanced Text Analysis TechniquesTopic Modeling