Research On Sentiment Analysis Method Of Comment Text Based On Big Data

With the rapid development of information technology and the popularization of the Internet,Internet users are more expressive about their viewpoint and attitudes on the Internet.Various online platforms store an unfathomable amount of text information,such as product reviews on shopping websites,comments on news media,comments on social networking sites,etc.Most of these text messages display user emotions.The sorting and analysis of these emotionally charged review texts on the Internet is of great benefit to various industries.Using the text sentiment analysis method in natural language processing can discover the sentiment tendency of Internet users from the comments they post.However,with the massive amount of text data on the Internet today,traditional sentiment analysis methods are not very effective,so this article designed a sentiment analysis model based on the big data method for these massive data.This article mainly introduces the research background and significance of sentiment analysis,and how to build a hybrid model based on MapReduce:(1)First use the Flume tool to extract a large amount of user comment data from Twitter,and preprocess these comment texts to remove data redundancy,errors and noise,such as some stop words,URLs,fuzzy words,and punctuations.The preprocessed data is stored in the Hadoop distributed file system as the experimental data for later stages.(2)Then the preprocessed comment data is transformed into word vectors.This paper proposes a hybrid word vector model of CBOW that is generated by mixing the CBOW word vector model,sentiment corpus and TF-IDF word weights.The word vector model is tested in contrast experiments,which demonstrated that the proposed word vector model can improve the F1 value of sentiment analysis.(3)Finally,the sentiment analysis method of the MapReduce hybrid model proposed in this paper is introduced.Comparative experiments are carried out in terms of the accuracy and speed of sentiment analysis.The experiment verifies that the proposed sentiment analysis method increased model speed while maintaining a high level of accuracy.In summary,this article tackles the weakness in big data processing among traditional sentiment analysis models.Using the CBOW hybrid model for word vectors and the MapReduce hybrid model for experiments can quickly and accurately implement sentiment analysis in a large-scale text database.It solves the time complexity problem of traditional methods remarkably.