Restaurant Consumer Behavior Analysis Based On BIGRU-AE Sentiment Model

With the rapid development of the Internet,the text information of user comments has been growing like a gushing well.In the era of Internet big data,the text data of user comments contains a lot of valuable information.The study of this paper involves the analysis of emotional polarity of text data.In recent years,the research in this direction has made good progress and shown a broad application prospect.This paper mainly studies the emotional analysis model of user comments,which is applicable to users and businesses in the field of restaurants,so as to accurately and intuitively obtain users' consumption behaviors and preferences for catering.The main content of the experiment is put forward based on BiGRU-AE sentiment analysis model,it is a two-way GRU helped study training methods,through the way of attention mechanism in terms of word vector and its weighted fusion sentiment analysis results,but also considers the Word2Vec and BERT two kinds of language model training process,using the word vector,vector,word vector is a combination of three different ways of data input,contrast the strengths and weaknesses of different input methods respectively.Through comparative experiments,it is verified that the model can improve the accuracy of text expression,and can consider the semantic characteristics and long-term dependence of text data.This paper has the following innovations:First,BiGRU-AE emotion analysis model is proposed to enhance the learning ability of emotion analysis in text context;Secondly,the GCAE model and SynATT model are improved.In addition,a word fusion model is proposed as data input in the process of data preprocessing.Emotional comments data analysis model is presented in this paper the basic thought and technology on the basis of the experimental system was designed based on the emotional analysis model of users food consumption behavior analysis experiment system,then has carried on the detailed design to model and verify the effect,according to the analysis of system requirements,regarding content requirements for design and implementation of the system.