The vast diffusion of devices equipped with a GPS receiver has brought the possibility of collecting data related to massive amounts of moving objects on a scale never seen before. During the latest years, such diffusion instigated the development of many different techniques to deal with location prediction problems. Existing works mainly aim at predicting the next location of moving objects by focusing on information in the spatial domain. In this paper we want to take into account information in the temporal domain as well, both to improve the reliability of predictions and to answer not only where a moving object is going to move, but also when an object is expected to leave its current location. To this end we propose TPRED, a framework based on probabilistic suffix trees which tries to capture typical movement patterns of moving objects, and computes reliable predictions accordingly, by exploiting information both in the spatial and temporal domains. In order to prove the validity of our contribution we conduct an extensive set of experimental evaluations, based on realworld datasets and different performance metrics, where we show the efficiency and effectiveness of our proposal.

TPRED: A Spatio-temporal location predictor framework

LETTICH, FRANCESCO;RAFFAETA', Alessandra;ORLANDO, Salvatore
2016-01-01

Abstract

The vast diffusion of devices equipped with a GPS receiver has brought the possibility of collecting data related to massive amounts of moving objects on a scale never seen before. During the latest years, such diffusion instigated the development of many different techniques to deal with location prediction problems. Existing works mainly aim at predicting the next location of moving objects by focusing on information in the spatial domain. In this paper we want to take into account information in the temporal domain as well, both to improve the reliability of predictions and to answer not only where a moving object is going to move, but also when an object is expected to leave its current location. To this end we propose TPRED, a framework based on probabilistic suffix trees which tries to capture typical movement patterns of moving objects, and computes reliable predictions accordingly, by exploiting information both in the spatial and temporal domains. In order to prove the validity of our contribution we conduct an extensive set of experimental evaluations, based on realworld datasets and different performance metrics, where we show the efficiency and effectiveness of our proposal.
2016
Proceedings of the 20th International Database Engineering & Applications Symposium, IDEAS 2016
File in questo prodotto:
File Dimensione Formato  
p34-TPRED.pdf

non disponibili

Descrizione: Articolo principale
Tipologia: Versione dell'editore
Licenza: Accesso chiuso-personale
Dimensione 536.84 kB
Formato Adobe PDF
536.84 kB Adobe PDF   Visualizza/Apri

I documenti in ARCA 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/10278/3680271
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 8
  • ???jsp.display-item.citation.isi??? ND
social impact