Research And System Implemen-Entation Of Aspect-Based Sentiment Analysis Algorithm For Restaurant Reviews

With the rapid development and popularity of the Internet,hundreds of millions of users access information and share their views on the Internet.Various types of user comment texts have grown exponentially,bringing a lot of valuable information.Text emotional tendency is the most important information,which can reflect the user’s real thought.With the development of the Internet,the restaurant industry has gradually changed the way it is served.Online business has become the main source of revenue for most restaurants,while generating a lot of user review data.The restaurant industry itself is a service industry,and its text sentiment analysis results have rich application scenarios,coupled with the availability of data,make the sentiment analysis task for restaurant reviews extremely valuable.Aspect-based sentiment analysis is also called attribute-level sentiment analysis,which is different from sentence-level sentiment analysis in that it analyzes specific attributes rather than the whole sentence.Aspect-based sentiment analysis tasks have more specific analysis objectives and are therefore more applicable.Based on ASAP,which is the largest Chinese restaurant review dataset,this paper studies the aspect-based sentiment analysis algorithm.On the basis of local context mechanism,the dynamic local context design is proposed to solve the core problem of how to find the context more relevant to the target attribute word.The experimental results of dynamic local context design presented in this paper are submitted to LUGE Platform for evaluation.The results show that dynamic local context design exceeds most of the schemes for aspect-based sentiment analysis,which proves the effectiveness of the algorithm design in this paper.Based on the results of aspect-based sentiment analysis algorithm,this thesis designs and implements a restaurant review analysis system.The system uses Vue framework to develop the front end and SpringBoot framework to build the back end.It realizes the development of functional modules such as system management,sentiment analysis and data management,and realizes online analysis and restaurant analysis functions in the sentiment analysis module to support users’ diverse experience of aspect-based sentiment analysis functions.