The relationship strength model between social network nodes is the key of social networks service such as information dissemination researches and recommendation system.The traditional researches focus on modeling simple binary relations and static relations, without considering dynamic attenuation of user interaction effects.Aiming at this problem, this paper proposes a social networks user relationship strength model based on Hawkes process(HP-URS), which takes the relationship strength, similarity and history interaction behavior between users as a latent factor, latent factor incentive and presentation respectively.This model uses Hawkes process to characterize relationship between history interaction behavior and user relationship strength.This model provides a solution of the disadvantages of the original model without considering user history interaction effects and their attenuation.This paper uses the data from microblog social networks evaluating HP-URS model, and the experimental results show that this model can improve relationship strength prediction accuracy and coverage rate of the Top-N neighbor nodes based on relationship strength.
[1] 荣辉桂,火生旭,胡春华,等.基于用户相似度的协同过滤推荐算法[J].通信学报,2014,35(2):16-24. Rong Hui-gui,Huo Sheng-xu,Hu Chun-hua,et al,User similarity based collaborating filtering recommendation algorithm[J].Journal on Communications,2014,35(2):16-24.(in Chinese)
[2] 徐志明,李栋,刘挺,等.微博用户的相似性度量及其应用[J].计算机学报,2014,37(1):207-218. Xu Zhi-ming,Li Dong,Liu Ting,et al,Measuring similarity between microblog users and its application[J].Chinese Journal of Computers,2014,37(1):207-218.(in Chinese)
[3] Xiang R,Neville J,Rogati M.Modeling relationship strength in online social networks[A].Proceedings of the 19th International Conference on World Wide Web[C].Raleigh,North Carolina,USA:ACM,2010.981-990.
[4] Zhang S.Influence of relationship strengths to network structures in social network[A].201414th International Symposium on Communications and Information Technologies[C].Incheon,South Korea:IEEE,2014.279-283.
[5] Gilbert E,Karahalios K.Predicting tie strength with social media[A].Proceedings of the SIGCHI Conference on Human Factors in Computing Systems[C].Monterey,California,USA:ACM,2009.211-220.
[6] Panovich K,Miller R,Karger D.Tie strength in question & answer on social network sites[A].Proceedings of the ACM 2012 Conference on Computer Supported Cooperative Work[C].Seattle,Washington,USA:ACM,2012.1057-1066.
[7] He Y,Zhang C,Ji Y.Principle features for tie strength estimation in micro-blog social network[A].2012 IEEE 12th International Conference on Computer and Information Technology[C].Chengdu,Sichuan,China:IEEE,2012.359-367.
[8] Liniger TJ.Multivariate Hawkes Processes[D].Zurich:Swiss Federal Institute of Technology Zurich,2009.
[9] Filimonov V,Bicchetti D,Maystre N,et al.Quantification of the high level of endogeneity and of structural regime shifts in commodity markets[J].Journal of International Money and Finance,2014,42:174-192.
[10] Bowsher C G.Modelling security market events in continuous time:Intensity based,multivariate point process models[J].Journal of Econometrics,2007,141(2):876-912.
[11] Hawkes A G,Adamopoulos L.Cluster models for earthquakes:Regional comparisons[J].Journal of the International Association for Mathematical Geology,1976,8(4):463-475.
[12] Juncen L,Sheng G,Yu Z,et al.Inferring links in cascade through Hawkes process based diffusion model[A].20144th IEEE International Conference on Network Infrastructure and Digital Content[C].Beijing,China:IEEE,2014.471-475.
[13] Ozaki T.Maximum likelihood estimation of Hawkes' self-exciting point processes[J].Annals of the Institute of Statistical Mathematics,1979,31(1):145-155.
[14] Guo C,Luk W.Accelerating maximum likelihood estimation for Hawkes point processes[A].201323rd International Conference on Field Programmable Logic and Applications[C].Porto,Portugal:IEEE,2013.1-6.
[15] 朱郁筱,吕琳媛.推荐系统评价指标综述[J].电子科技大学学报,2012,41(2):163-175. Zhu Yu-xiao,Lv Lin-yuan.Evaluation metrics for recommender systems[J].Journal of Electronic Science and Technology.2012,41(2):163-175.(in Chinese)