Research On Mobile Service Recommendation Algorithm And Model Based On User Context

With the rapid development of mobile Internet,social network,electronic commerce and other related technologies,mobile services and their information content have increased dramatically.This kind of phenomenon brings convenience to people,but it also brings people the problem of " Mobile information overload" at the same time.This problem not only affects the user's use experience,but also reduces the utilization of network resources.Therefore,the academia and the industry regard this as a hot issue at present.Therefore,it is of great significance to explore how to personalize the recommended system and let the user have a better use of the experience.This paper studies the algorithm and model of mobile service recommendation.The main contents of this paper are as follows:(1)To solve the problem of sparsity and accuracy of mobile Internet in mobile Internet,a mobile service recommendation algorithm based on user behavior and context-aware is proposed.The algorithm first obtains the information of the target user's context,such as time,location and so on,and then filters similar users' behaviors in similar situation.Secondly,using the behavior of mobile users(browse,collection,purchase,etc.)to obtain the target user's demand for a service.Use this to make up the user rating matrix,and then calculate the similarity between the users according to the matrix.Finally,we calculate the preference of the target user for various services in this situation in order to generate the recommendation.The experimental results show that MSR-BCA algorithm can calculate user's demand for services by using the behavior of mobile users,which effectively alleviates the problem of sparsity of score data.At the same time,the combination of situational information has a better recommendation effect.(2)In view of the lack of formal logic research in the field of service recommendation,a mobile service recommendation model based on situational calculus is proposed.First,on the basis of analyzing the basic methods of mobile service recommendation,the logical theory system of scenario calculus is introduced.In the recommendation system,user behavior acquisition,data matrix construction and reconstruction,data access and extraction,similarity and prediction score calculation,and service recommendation to users are considered as action domain.Also,the effect that is used to describe the above action is regarded as fluent domain.At the same time,the flow of changes in the mobile service recommendation process is considered as a situation domain.Furthermore,the mobile service recommendation process is defined as the process of taking the initial situation as the starting point and evolving dynamically according to the rules which development in the action domain,the fluent domain and the scene domain.On this basis,the MSR-SC action action axiom is built on the basis of the general process of service recommendation,the capture situation and the law of situation change.Finally,in order to verify the logical validity of the model,a formal description of the mobile service recommendation example using MSR-SC is given.According to the description example,the model is correct and effective.(3)The mobile service recommendation platform is built on the basis of the research of the foregoing algorithms and models.In this platform system,the problem of data sparsity and accuracy in platform engineering is solved by using the MSR-BCA algorithm in this paper.The MSR-SC scenario model is used to make up for the logic completeness in the platform engineering.It can be seen from the engineering practice that the platform has a good recommendation effect in the case of sparse data.