Research On Online Health Community Information Topic Features And Sentiment Classification Based On LDA Model

Under the environment of ‘Internet plus health care’,online health community,as a new platform for modern medical information acquisition and interaction,has gradually become a key node for health information interconnection in the medical industry.The online health community is in the innovation stage of the online medical service industry.The information data of the community is huge,complex and heterogeneous,and multi-source.The emotional needs of users are increasingly urgent.Therefore,it is significant to accurately grasp the information interaction theme of online health community,study the information theme characteristics of the community and its change trend,analyze the emotional appeals among users,and further explore the potential information value,which is vitally important to optimize the classification and management of information theme of online health community,and is helpful to implement the policy of ‘Internet plus health care’ and improve the service quality of online diagnosis and treatment of the public.This study analyzes the theme characteristics of the online health community of Baidu Gout Post Bar,builds a theme classification and emotion classification model,and explores the information theme characteristics and emotional preferences of community users.First of all,through the design of multi-thread crawler code,crawl the text data of online health community posts and postings,and get the basic text corpus of online health community after data cleaning and preprocessing.Secondly,the LDA theme feature model is built based on the corpus,and the distribution and evolution trend of the theme features of online health communities are explored using different visualization methods.Finally,machine learning method is used to classify the emotion of the text corpus,and the corresponding information subject feature results are studied for the positive and negative corpus respectively.The research results show that the online health community information includes five thematic feature categories: lifestyle,drug treatment,disease diagnosis and treatment,pathological knowledge and emotional support.The hot topics of community user information interaction form an information interaction pattern with drug treatment as the main body,and the other four thematic categories are coordinated and parallel.Among them,the demand for medical treatment information shows a slow upward trend,the demand for disease diagnosis and treatment and lifestyle information shows a downward trend,and the demand for emotional support and pathological knowledge information tends to be stable after a small fluctuation.In addition,the emotional classification of online health community information is mainly based on neutral emotional data,followed by positive emotional data and negative emotional data.Finally,according to the research results,the following three suggestions are put forward:Firstly,deepen the information system construction of online health communities.Secondly,pay attention to and follow up the emotional needs of community users in real time.Thirdly,improve the medical system for chronic diseases.