User Comment Sentiment Analysis Based On Big Data

With the rapid development of e-commerce and social media,netizens have formed a habit of expressing their own opinions or opinions about various transaction objects or subjects that are closely related.It has created a huge amounts of user reviews data.Reviews data is the most comprehensive way to fully reflect the public opinion of the largest network users and the personal needs of the Internet users themselves.The traditional emotional analysis method has been unable to comprehensively and effectively analyze the data of massive the reviews.Therefore,this paper makes a sentiment analysis for massive user reviews,proposes a sentiment analysis model for user reviews based on convolutional neural networks,and constructs an emotional modeling method based on the OCC sentiment model to implement sentiment analysis of user reviews.The main research contents of this thesis are as follows:Research on Emotion Modeling Method Based on OCC Model.This thesis first compares the classic Hidden Markov Model(HMM)and the OCC model,and analyzes the performance and applicable scope of the two emotion modeling methods.This thesis proposes the use of OCC model for emotional modeling,and based on the OCC model.Defining 6 types of emotions,combining the natural language processing tools and HowNet dictionary to carry out emotional rec.ognition and emotional annotation on user reviews.Using the user comment sentiment analysis model to perform an experiment on the emotionally annotated training set,the result shows that the training set using the OCC model emotionally annotated the accuracy rate of the user comment sentiment analysis model.Based on the emotional modeling based on the OCC sentiment model,the user review sentiment analysis model is studied.Firstly,the comparative study of the classic machine learning model can not fully analyze the emotions of the massive user comments.This paper proposes the use of convolutional neural network model to carry out sentiment analysis of user reviews.Firstly,the training set of user review data is preprocessed,then the user review data is vectorized through the Word2vec model,and on this basis,a convolutional neural network is used for sentiment analysis.This thesis is also in e-commerce,social networking,and networking respectively.In the public opinion field,fine-grained examples are given for analysis,and the effectiveness and practicality of the model are verified through experiments.