Learning Significant Locations and Predicting User Movement with GPS
ISWC '02 Proceedings of the 6th IEEE International Symposium on Wearable Computers
A suffix tree approach to anti-spam email filtering
Machine Learning
Evaluating Variable-Length Markov Chain Models for Analysis of User Web Navigation Sessions
IEEE Transactions on Knowledge and Data Engineering
Understanding mobility based on GPS data
UbiComp '08 Proceedings of the 10th international conference on Ubiquitous computing
NextPlace: a spatio-temporal prediction framework for pervasive systems
Pervasive'11 Proceedings of the 9th international conference on Pervasive computing
A universal predictor based on pattern matching
IEEE Transactions on Information Theory
Hi-index | 0.00 |
Models developed for the prediction of location, where a specific individual will be present at a future time, are typically implemented using a one-model-per-user approach which cannot be employed for inferring collective or social behaviours involving other individuals. In this paper, we propose an alternative that allows for inference though a collaborative mechanism which does not require the profiling of individual users. This alternative utilises a suffix tree as its core underlying data structure, where predictions are computed over an aggregate record of behaviours of all users. We evaluate the performance of our model on the Nokia Mobile Data Collection Campaign data set and find that the collective approach performs well compared to individual user models. We also find that the commonly used Hit and Miss score on its own does not provide sufficient indication of prediction accuracy, and that employing additional metrics using the mean error may be preferable.