The LS-CONTEXT project investigates principled methods to extract human behavioral patterns from longitudinal cell phone data.
Cell phones are increasingly equipped with accelerometers, GPS, Bluetooth, and other sensors, and can ubiquitously sense daily life activities for large populations, opening novel possibilities for research in several sciences and for mobile applications. The project is developing methods to represent human behavior based on the integration of the various sensor data types available on cell phones, and developing machine learning methods to mine long-term personal and group routines in terms of mobility, proximity, and communication. The project also aims to create research resources for behavioral modeling from mobile phones used in real life.
The research in the project is funded by and done in collaboration with Nokia Research Lausanne.