Method for establishing dynamic instant recommendation service model under mobile spatio-temporal awareness

The invention discloses a method for establishing a dynamic instant recommendation service model under mobile spatio-temporal awareness. The method comprises the steps that a collected mobile user context log data set is preprocessed; a CIT-LDA theme model is used for extracting mobile spatio-temporal awareness information and user behavior preference change information under each time slice for semantic modeling; the drifting probability of the interested theme of a user is calculated through symmetric KL divergence, and an interest change point of the user is determined; evolution of a userinterest sub-theme drifting track is divided into three types; according to long-term interests, short-term interests and weight of each interest of the user, a user behavior preference model is determined; the personalized dynamic instant recommendation service model based on Top-K is determined. In a Hadoop parallel distributed computing environment, the CIT-LDA theme model, and exploratory research of personalized reco