Research And Application Of Real-time Interest Model Based On Context-Aware

With the development and progress of the times,the variety of information are explosive increasing.Those information not only provide convenience to our life,but also made problems about information overloaded.As we know that one's energy is limited,can not quickly select the thing which they were interested from hundreds of millions of information.Recommendation System can provide people very specific and targeted imformation.Therefore,the research of Recommendation System become more and more valuable.According to the users' behaviors,registration informations,browsing history.Recommendation System can pick out the things that the users are truly interested.However the traditional Information Recommendation System are based on the connection of user to user,user to their programs.About the situation of the users' being were hardly any weight.And the system are took to much outdated information,thereby the system cannot make users receive real-time feedback.causing the final recommends are far away form the excepted goal.In this paper based on the analysis of the Context Awareness and Date Modeling Theory,and compared with the traditional Information Recommendation System to get a solution to improve a real-time user model.Based on situation improved,and the integration of user interest,label behavior,users will be the Situational Factors in the environment into the interest model,and with the situation of Mining User Association between the accurate access to changing user preferences,in order to facilitate the recommendation system to recommend for users meeting the user preferences of the project.Finally,the improved interest model is evaluated on the small data set of watercress.The experimental result shows that the improved user interest model can more accurately reflect the user's real-time interest.On the basis of a certain recommendation algorithm,the model is applied to the social hot spot of "missing children",and an example is analyzed.This article mainly carries on the thorough research in the following aspects:First of all,context aware recommendation algorithm and related theory knowledge,the specific definition of the situation,the effects on the different situations to the user interest,and user context classification,according to the situation given access to the specific identification method.At the same time,compared with three kinds of the traditional Personalized Recommendation Model--content(based recommendation,knowledge based on recommendation and collaborative filtering recommendation)based on a detailed description,respectively analyzes three types of recommended mode has advantages and disadvantages.Secondly,to the point of the traditional user interest model was very bad at dynamic response ability and the less influential factors of interest,in this paper explores the user's situational interest into the user's interest model.First of all to the users' interest label are standardized,effectively solve the semantic problem of tags,stabled the label design to the user interest model;secondly,according to interest from Internet users' browse,search,download and access to the user's behavior,to establish user interest model of real-time behavior;then,by obtaining recognition for user context RFID/GPS technology and situation model,establish user situational interest model;finally,the integration of the three model get a real-time user model based on context.This model can not only reflect the real interest of users more comprehensive and accurate,and can be updated automatically according to the users in the use of the experience in the process of data model,which can guarantee the real-time response to changes in the user's interest,so that personalized recommendation system can be more accurate for customer service.Finally,about the problem of missing children,everyday there are numerous of children missing information.Such massive information can be very likely cost the data overload,And the useful information are very hardly to find.And delay to find those missing child,therefore,proposed the recommendation system of the user interest model to real-time information about missing children,based in certain recommendation algorithm,accurately recommend to users that they are interested in and concerned about missing children information,in order to save the user's time and energy,improving the chances of children back.