The Research And Application Of Wechat Official Accounts Information Mining In Enterprise Information Service

With the continuous development of mobile Internet technology and smart phone,a large amount of information data is generated every moment.We Chat is one of the most widely used applications,with hundreds of millions of users,and the number of We Chat Official Accounts has been increasing in the past two years and contains a lot of valuable information.How to make good use of this data repository is a hot topic in the current study.In this paper,we first crawl the content of We Chat Official Accounts,and then preprocess the resulting text data to establish a topic mining model,using the two topic models of LDA and HDP clustering analysis of the text data.The two algorithms are studied and compared.At last,the Official Accounts information recommended platform is designed and implemented for enterprise users.This platform allows users to obtain information about the Official Accounts they are interested in.The specific work of this paper includes the following aspects:1.A new method for data fetching of We Chat Official Account is proposed.The We Chat Official Accounts crawler is implemented through android emulators,automatic click APP,packet resolution and storage.The actual test shows that the crawler has the characteristics of high efficiency,high real-time performance,high data integrity and high stability,which is one of the innovations of this research.2.Based on the keyword extraction method of Jieba segmentation tools and Latent Dirichlet Allocation(LDA)and Hierarchical Dirichlet Process(HDP)topic model method to We Chat Official Account content data mining to form a category,and then extract the keywords from these categories to form the tags.The influence of LDA and HDP model parameters on text clustering analysis results is to be researched.The accuracy rate,recall rate and F1 value of the two algorithms were analyzed and compared by experimental data.3.The design and implementation of the Official Accounts information recommendation platform for enterprise users.The main functions of the system are:(1)Provide users with a list of enterprise labels to choose.(2)The platform pushes the information about the articles of interest to the user through the recommendation algorithm according to the user's selectedTags.(3)The platform will provide users with comprehensive Official Accounts content information,including articles,reviews,readings,release time,Official Account names,etc.This paper builds up the real-time information service platform for enterprises through efficient and stable Official Accounts data crawling and accurate text topic classification.The related content has been applied in the laboratory,and it has certain innovation and good practical application prospect.