| Nowadays,with the development of the Internet is more mobile,and more intelligent,mobile products,users can access to vast amounts of data through mobile intelligent devices.The data contains mobile user's interests,preferences and behavior trajectory information,and has great value of application value.This topic mainly research the heterogeneous scenario source under the mobile users' personalized recommended,focuses on the user data preprocessing,and the prediction results applied to the user scene sequence extraction in the study,mainly done the following work:(1)Lack of user scenarios sequence information filled is one of the key problem in the research of data preprocessing in multi source data fusion.Aiming at the lack of information users in the scene data sequence problem,the problem of the missing in Rolling Grey Model(Rolling Grey Model,RGM)in this paper,on the basis of an improved method is proposed.The method of user stories known state power refers to transform the original data sequence,by transform the data after rolling grey prediction and to predict the data power refers to the inverse transformation,thus draws the scene state missing data prediction.Experiment shows that this method can improve the accuracy of the forecast data to a great extent.(2)Due to the diversification of mobile intelligent terminal sensors,the relationship between the scene source data information has become more and more closely.In view of this situation,scene source user context sequence extraction based on Hidden Markov Model(Hidden Markov Model,HMM)improved model is proposed.Model is first through the mobile sensors to collect user context information,using the Cartesian product of scene source data information fusion,finally the HMM model used to extract the mobile user data sequence.The experimental results show that the proposed method can effectively extract the user's scenario data sequence for the recommendation of service research,and has obtained the good effect. |
PDF Full Text Request