Review Based Local Experts Discovery In LBSN

The development and popularization of online social networks have brought great convenience to people's daily lives.Today,there are billions of users being active in social networks,resulting in lots of social information.As a huge platform with information exchange and sharing of life service,it has a profound impact on people's way of life,such as facing problems,users will ask for the expert in social network,so experts identification is particularly important.However,the demand for experts is not only the areas of expertise,but.the location factors,which is so called local experts.Local experts can provide users with more personalized service and have important implications for the collection and dissemination of local information.Location-based social network(LBSN)fuse the online relationship and the offline behavior.It has built bridges between the virtual social space and the real behavior through location characteristics and its mass data offers an important theoretical basis for this study.This paper analyzes the topic authority and location authority of individuals,and then quantifies their professional degree.Based on the LBSN data,a local experts discovery algorithm(LER)is proposed with fully using of all kinds of information in the network.Firstly,the regional characteristics and category preferences of users are analyzed to confirm the objective existence of local experts in the network.And then a set of candidates is formed based on the given query.In order to measure the level of candidates more accurate,this article scores candidates from personal attributes,friends relations,review semantics and location,and the local expert scoring model is constructed by linearly adding the multidimensional scores.In order to facilitate the study of the model,this paper selects the POI in the city based on clustering algorithm,and the local experts are marked.At the same time,the effective algorithm based on maximal likelihood estimation is design to train and evaluate the score weights.Finally,the local expert discovery algorithm is proposed,and the prototype system is implement.In order to verify the effect of the proposed algorithm,this paper has carried on experiment using several expert discovery algorithms on the real dataset from the review social network Yelp.Through the comparison and analysis of the experimental results,we can get our algorithm(LER)achieves a better effect.