Design And Implementation Of User Interest Model In Wireless City Repository

As the application of Smart City (SC) in China, Wireless City (WC) is proposed to solve the problem of citizen life informatization. With the function of processing data into information and providing support for the application, the core task of WC is the collection, integration and analysis of large amount of data. Meanwhile, the solution of this task cannot be separated from the audiences which we also named users. How to avoid the rising problem of information overload and transform the data to information which is practical significative as well as human interested is the burning issue of WC.Be responsible for model storage, model mapping and model association analysis of the big data of WC, Wireless City Repository (WCR) which is the core of WC is based on knowledge engineering and knowledge repository. By means of exploring the correlation between the information, WCR sets target of the regularization of unordered set of data, meeting the individual needs of the user and aggregating information as maximum as possible. In short, WCR focus on the whole flows of data collection, data extraction, data analysis, data modeling, data processing and user recommendation.WCR’s core concepts are information and user. Aimed at reducing the participation of user, analyzing the behavior flow of user and exploring the interest of user, User Interest Model (UIM) is the emphasis of this thesis. Different from traditional UIM, the UIM of WCR pays close attention to big data and various categories. By processing the unordered set of big data in WCR and analyzing the information associated with user, implement the model recording and expression of user interest. Meanwhile, the recommended system and search engine which are the service based on UIM is in charge of recommending interested results for users. These results verify the feasibility and efficiency of the UIM which will amend and update the UIM better.The principal contributions of the work presented in the thesis are:(1) Design idea as well as implementation proposal is proposed based on analysis of the actual demands and application scenarios of WC as well as summarizing the data features of WC.(2) Process flow of WC is designed according to analysis of the advantages and disadvantages of the traditional user interest model as well as characteristics of WCR. (3) A conceptual UIM for WCR is proposed, which is divided into User Information Relevance Model, User Content Relevance Model and User Behavior Relevance Model according to mathematical modeling, statistics, data mining, natural language process and graph theory. This conceptual UIM can be implemented within the technology of model fusion.(4) Automated update technology of UIM is discussed and proposed. Service based on UIM is proposed according to recommended system and search engine as well as consideration of cold start problem.(5) Versatility and extension method of UIM for WCR is discussed as well as the suggestion on how to use UIM in other system is put forward.(6) Experimental results including test methods and evaluation indicators based on the UIM proposed in this thesis is showed compared to the traditional user interest model. These results imply the superiority of the present method.