The pervasiveness of mobile devices and location based services produces as side effects an increasing volume of mobility data which in turn create the opportunity for a novel generation of analysis methods of movements behaviors. In this paper, we propose a method Where Next aimed at predicting with a certain accuracy the next location of a moving object. The prediction uses previously extracted movement patterns named Trajectory Pattern which are a concise representation of behaviors of moving objects as sequences of regions frequently visited with typical travel time. A decision tree, named T-pattern Tree, is built and evaluated with a formal training and test process. Using Trajectory Patterns as predictive rules has the following implications: (I) the learning depends by the movement of all available objects in a certain area instead by the individual history of an object; (II) the prediction tree intrinsically contains the spatiotemporal properties emerged from the data and this allows to define matching methods strongly depending on such movement properties. Finally an exhaustive set of experiments and results on the real dataset are presented.

Location prediction through Trajectory pattern mining (extended abstract)

Pinelli F.;
2010-01-01

Abstract

The pervasiveness of mobile devices and location based services produces as side effects an increasing volume of mobility data which in turn create the opportunity for a novel generation of analysis methods of movements behaviors. In this paper, we propose a method Where Next aimed at predicting with a certain accuracy the next location of a moving object. The prediction uses previously extracted movement patterns named Trajectory Pattern which are a concise representation of behaviors of moving objects as sequences of regions frequently visited with typical travel time. A decision tree, named T-pattern Tree, is built and evaluated with a formal training and test process. Using Trajectory Patterns as predictive rules has the following implications: (I) the learning depends by the movement of all available objects in a certain area instead by the individual history of an object; (II) the prediction tree intrinsically contains the spatiotemporal properties emerged from the data and this allows to define matching methods strongly depending on such movement properties. Finally an exhaustive set of experiments and results on the real dataset are presented.
2010
Prediction
Spatial and temporal mining
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11771/23919
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
social impact