A Theme-Context Mixture Model for Personalized Search in Social Network
Dongling Chen, Wen Hua Zeng · 2015
Nowadays, social network technology provided a lot of ways for users to express their emotions and attitudes online.How to model user preferenced information and provide personalized service is a crucial problem in big data era.In this paper, a new probabilistic model be proposed to model and analysis topic trends in personalized search.The model extended the Latent Dirichlet Allocation (LDA) model by introducing context variables, through which we can detect and analysis topic trends according to contextual information.The core idea of proposed probabilistic model is to learn a finite Dirichlet mixture model, and then adopt Bayesian discriminant to detect topic and topic trends analysis.Experimental results show that the proposed probabilistic mixture model can detect topics and discover topic trends effectively.