Modeling Research Of User's Dynamic Interests Based On Ontology And Folksonomy

As the 43 th figures on “China Internet Development Statistics Report” show,our country web cam scale expands to 8.29 billion by December 2018,and the Internet penetration rate reached 59.6%.Along with the personal development of online reading platforms,more and more customers are enjoying the convenience of cyber reading,and becoming increasingly dependent upon the recommendation of book resources.With the guide of consumption pattern of quick-hit reading,appropriate recommendation of related resources helps users' information consumption becoming much more time-saving and cost-saving,and improves users' viscosity of reading platform,which also tests the ability of application manufacturers and product designers to understand users' needs.How to provide accurate and real-time resource recommendations for users by their interactive behavior of the internet application platforms is a hot research topic in information recommendation,information retrieval,information quality,user's behavior,user's interest and so on.Social tagging is the product of Web 2.0,which is users' initiative behavior records including marks,groups and classification on the information resources.In order to better satisfy customers' demands of individuality reading,these application platforms which are similar with Douban reading are helping people enjoying recommendation of related resources by users' labeling information.Social tags reflect users' interest information,and it's the common research method of information recommendation on folksonomy systems to get users' interest and preference by rational data analysis algorithms or models.It was noted that users' interests and preferences were influenced by seasonal variations,environmental changes,ages and experience in real life as time went on.However,some researchers in related fields have not considered the influence of dynamic change of customers' interests in folksonomy systems on accuracy and validity of folksonomy recommendation algorithms or models,which makes most of analysis methods apply only to the resource recommendation of users' long-term or stable interests,and can not meet the current demand for great efficiency and accuracy of information services.Therefore,the design of interest models and analysis methods of related information data that adapting to the dynamic change of social label users' interests are important measures for online reading service at the present situation,which are also urgentproblems to be solved in the research and application of information recommendation,information retrieval and information quality based on social labels.Nowadays,there are few research on mining user's dynamic interests based on ontology and folksonomy.From this angle,the paper mainly uses literature analysis,comparative analysis,model constructing and empirical research methods.On the basis of the existing ontological constraints and relationship definitions of words in Chinese Classification Thesaurus and Chinese Library Classification,a tags-based ontology was constructed by using FCA,which reflects the Douban sample's interest.The results of forecast experimental on interest intensity index and stability index indicate the suitable representation of interest index on tags.On these foundation,the paper proposes the user's initial interest model,and the renewal process of interest nodes,which is mainly based on interest intensity and supplemented by interest stability index.The paper proposes an interest model reflection user's dynamic interest characteristics by combining the research methods of ontology analysis and model construction of user's interest,and it will summarize the methods of model updating according to the dynamic change characteristics of user's interest,so as to provide reference for the research or application on information recommendation,information retrieval,information quality and other related fields.