Reality Mining is comprised of communication and proximity data of 97 faculty, student, and staff at MIT recorded continuously via preinstalled software on their mobile devices over 50 weeks [Eagle et al. 2009]. From the raw data we built sequences of weekly temporal graphs for three types of relations; voice calls, short messages, and bluetooth scans. For voice call and short message graphs a directed edge denotes an incoming/outgoing call or message, and for bluetooth graphs an edge depicts physical proximity between two subjects. The ground truth captures semester breaks, exam and sponsor weeks, and holidays.
Less is More: Building Selective Anomaly Ensemble with Application to Event Detection in Temporal Graphs. Shebuti Rayana, Leman Akoglu, SIAM SDM, Vancouver, BC, Canada, April 2015
Less is More: Building Selective Anomaly Ensemble. Shebuti Rayana, Leman Akoglu, Transactions on Knowledge Discovery from Data (TKDD), May, 2016
Description: Date file contains the dates, Source file contains the Source ID, and Destination file contains the Destination ID for different types of communications. GroundTruth file contains the ground truth information.