Predicting next location of Twitter users for surveillance


Gunduz S. , YAVANOĞLU U. , SAĞIROĞLU Ş.

12th International Conference on Machine Learning and Applications (ICMLA), Florida, Amerika Birleşik Devletleri, 4 - 07 Aralık 2013, ss.267-273 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası:
  • Doi Numarası: 10.1109/icmla.2013.134
  • Basıldığı Şehir: Florida
  • Basıldığı Ülke: Amerika Birleşik Devletleri
  • Sayfa Sayıları: ss.267-273

Özet

In this study a novel approach that uses location based social networks for next location prediction in the field of technical surveillance and digital forensics is proposed. With the help of proposed methodology, search area for the potential criminals will be narrowed so that the spent time, money and effort by the law enforcement officers will be minimized. After collecting enough past location information for Foursquare users, the whole data is trained by means of Artificial Neural Networks. After training process, predicting the next location of the wantedperson is carried out. Prediction process is made region-based, so it is tried to predict the region of the potential criminals' next geographical location. The experimental results have shown that the proposed approach and developed system might achieve the prediction goal with only 3% error rate, and proposed methodology can be used by law enforcement officers for forensic surveillance and similar criminal acts.