Predictability of Human Behavior using Mobility and Rich Social Data

Ana Martic

AbstractThis thesis explores how predictable human mobility is, and whether knowing about mobility patterns of other people, who visit same places, can contribute to better prediction results. Human movements are periodical to some extends, which means that it is possible to create a model which can predict next place of a person in some moment based on the data about previous person's movements. In this thesis, an ensemble method is adopted, which gathers predictive power of multiple models, each capturing different human mobility features. The predictive models are build using GPS data collected for 136 experiment participants, during seven and a half months period. Prior to predictive modeling, data was carefully preprocessed and characteristics of human mobility are analyzed using multiple visualization techniques.
TypeMaster's thesis [Academic thesis]
Year2013
PublisherTechnical University of Denmark, DTU Compute, E-mail: compute@compute.dtu.dk
AddressMatematiktorvet, Building 303-B, DK-2800 Kgs. Lyngby, Denmark
SeriesM.Sc.-2013-45
NoteDTU supervisors: Jakob Eg Larsen and Sune Lehmann Jørgensen, DTU Compute
Electronic version(s)[pdf]
Publication linkhttp://www.compute.dtu.dk/english
BibTeX data [bibtex]
IMM Group(s)Intelligent Signal Processing


Back  ::  IMM Publications