Output details
11 - Computer Science and Informatics
Goldsmiths' College
Discovering routines from large-scale human locations using probabilistic topic models
<15> This journal is the 8th highest ranked in computer science, according to SCImago, with a 2012 impact factor of 5.576. The paper proposes a novel framework for discovering mobility patterns in mobile phone data, applying sophisticated unsupervised machine learning models to noisy real data. Methods based on feature visualisation (e.g. days, user types, locations) were used for validation. The paper has been widely cited, including by Facebook researchers in a PAMI (SCImago highest ranking CS journal) article. The paper directly led to an invitation to Farrahi to work with Sandy Pentland at the MIT Media Lab.