Normalized Cuts and Image Segmentation
IEEE Transactions on Pattern Analysis and Machine Intelligence
Using GPS to learn significant locations and predict movement across multiple users
Personal and Ubiquitous Computing
From context to content: leveraging context to infer media metadata
Proceedings of the 12th annual ACM international conference on Multimedia
Reality mining: sensing complex social systems
Personal and Ubiquitous Computing
Factorial Switching Linear Dynamical Systems Applied to Physiological Condition Monitoring
IEEE Transactions on Pattern Analysis and Machine Intelligence
Learning and inferring transportation routines
AAAI'04 Proceedings of the 19th national conference on Artifical intelligence
Place lab: device positioning using radio beacons in the wild
PERVASIVE'05 Proceedings of the Third international conference on Pervasive Computing
Accurate GSM indoor localization
UbiComp'05 Proceedings of the 7th international conference on Ubiquitous Computing
Learning and recognizing the places we go
UbiComp'05 Proceedings of the 7th international conference on Ubiquitous Computing
Mobility detection using everyday GSM traces
UbiComp'06 Proceedings of the 8th international conference on Ubiquitous Computing
Practical metropolitan-scale positioning for GSM phones
UbiComp'06 Proceedings of the 8th international conference on Ubiquitous Computing
AccuLoc: practical localization of performance measurements in 3G networks
MobiSys '11 Proceedings of the 9th international conference on Mobile systems, applications, and services
Improving activity recognition without sensor data: a comparison study of time use surveys
Proceedings of the 4th Augmented Human International Conference
A survey on smartphone-based systems for opportunistic user context recognition
ACM Computing Surveys (CSUR)
Hi-index | 0.00 |
This paper presents novel methodologies for the analysis of continuous cellular tower data from 215 randomly sampled subjects in a major urban city. We demonstrate the potential of existing community detection methodologies to identify salient locations based on the network generated by tower transitions. The tower groupings from these unsupervised clustering techniques are subsequently validated using data from Bluetooth beacons placed in the homes of the subjects. We then use these inferred locations as states within several dynamic Bayesian networks (DBNs) to predict dwell times within locations and each subject's subsequent movements with over 90% accuracy. We also introduce the X-Factor model, a DBN with a latent variable corresponding to abnormal behavior. By calculating the entropy of the learned X-Factor model parameters, we find there are individuals across demographics who have a wide range of routine in their daily behavior. We conclude with a description of extensions for this model, such as incorporating contextual and temporal variables already being logged by the phones.