Urban Anomalous Events Analysis Based on Bayes Probabilistic Model from Mobile Phone Records
We present an approach to detecting and analyzing urban anomalous events by Bayes Probabilistic Model. Using actual mobile phone data, we compute individual probability and get individual anomalous index under comparing occurrence probability and ordinary probability in a certain region and period. Expanding individual analysis to group analysis, we make statistics on anomalous activities of group and get their regularity so that we can measure the degree of deviation among activities of group during certain period and the regularity and finally judge whether urban anomalous events take place. Taking two areas in Kuming city, China as case study, we demonstrate effectiveness of our approach.
KeywordsMobile phone data Bayes probabilistic model Anomalous event analysis Individual anomalous analysis Group anomalous analysis
This work is supported by National Nature Science Foundation of China under grant no. 41231171. The authors would like to thank Xiaoqing Zou at Kunming University of Science and Technology, Kunming, China for providing us with mobile phone data.
- 1.Rojas, F., Calabrese, F., Fiore, F.D., Krishnan, S., Ratti, C.: Real Time Rome. MIT Senseable City Laboratory, Cambridge (2006)Google Scholar
- 7.Zang, H., Baccelli, F., Bolot, J.: Bayesian inference for localization in cellular networks. In: Proceedings of IEEE INFOCOM, San Diego, California, USA, pp. 1–9. IEEE Press (2010)Google Scholar