Talk – Leveraging Periodicity in Human Mobility for Next Place Prediction
May 29 all-day

In cooperation with IEEE Bangalore Section and TI India Technical University

IEEE CAS Bangalore Chapter is pleased to announce a seminar

Speaker: Bhakar Prabhala, Penn State University

May 29, 2014

11.00am – 12.00 noon

Venue – Texas Instruments, Bagmane Tech Park, CV Raman Nagar, Bangalore 560093


This seminar is open to IEEE members and non-members. There is no registration fee to attend this seminar, but you must confirm your participation by writing to before May 27. Your participation will be confirmed and further instructions will be sent to you.

Periodic transitions from place to place are inherent in human movements. Through visual examination we detect these periodic movements in traces of user tracking data. However such user tracking data sets tend to be sparse and incomplete. In addition, periodic movements are surrounded by noise: transitions to and from less frequently visited places and transitions to one of a kind visits. We present algorithms leveraging techniques and models to detect periodicity in individual user movements. Our algorithms predict a user’s next place given only the current context of time-stamp and location. We apply these algorithms to real user mobility data sets. Prediction accuracy depends on the ratio of periodic movements to noise in user traces.

We trace our submission to Nokia’s Mobility Data Challenge for next place prediction task, which took the second place. Aggregated average accuracy of predictions across all users is the sole criteria for the Challenge. We consolidated the techniques from our submission to two algorithms: PeriodicaB and PeriodicaS. We apply these algorithms to MDC data set as well as the WTD data set from UCSD study. We analyze accuracy of predictions for individual users as well as the aggregated average accuracy across all users. These results and the analysis were presented at the IEEE Wireless Communications and Networking Conference 2014. Current research aims to generate end to end trajectories for users in the trace data sets using periodicity and place semantics. By overlaying these trajectories we can determine intersection points that result in models for opportunistic mobile networks.

Bhaskar Prabhala is a researcher in the Institute for Networking and Security Research( at Penn State. His research interests include modeling, mobility prediction, and communication services in opportunistic mobile networks. Before returning to Penn State, Bhaskar set up the R&D Center for Royal Philips/NXP Semiconductors in Shanghai, China in 2006 and led the center as the General Manager until 2009. Before then, he was the General Manager of R&D Center (San Jose, California) in Philips Semiconductors for five years. Previously, Bhaskar was vice president of engineering at Syntax (acquired by LSI Logic in December 2000), a leading provider of advanced server solutions for corporations requiring cross-platform access to mission critical information. He was at Sun Microsystems for eight years in software development management positions and as a director of networking software. He moved to Silicon Valley in 1979 and joined Intel as a member of the 286 processor software architecture team. He was part of the faculty in Computer Science Department at Indiana University (Bloomington) from 1977 to 1979. Bhaskar graduated from Pennsylvania State University with an M.S. in Computer Science.