City Research Online

Exploring mobile trajectories: An investigation of individual spatial behaviour and geographic filters for information retrieval - Volume 2

Mountain, D. M. (2005). Exploring mobile trajectories: An investigation of individual spatial behaviour and geographic filters for information retrieval - Volume 2. (Unpublished Doctoral thesis, City, University of London)


This section presents both quantitative results, related to the performance of prediction surfaces according to the evaluation criteria described in section 3.5, and qualitative results, based upon testing geographic filters for information retrieval in an outdoor mobile computing environment as part of a major user evaluation study in the Swiss National Park, in the Summer of 2004.

The chapter begins with quantitative evaluation, where three test scenarios are described (walking, driving and daily migration), followed by a description of the systematic variation in the temporal component of prediction, designed to compare short-term (10 minutes) and long-term (60 minutes) predictions into future. Next, three geographic filters described in the methodology are introduced as prediction surfaces. The three filters are spatial proximity, temporal proximity and speed-heading predictions. For each approach, the prediction input parameters were varied in a systematic way to uncover the impact of buffer size and distance decay functions (for spatial proximity prediction surfaces), the ‘recent behaviour period’, temporal weighting and decay function (for speed-heading prediction surfaces) and the time budget and enclosing function (for temporal proximity prediction surfaces). Next, the results are presented and analysed to describe, explain and contrast the characteristics of each prediction approach. This is followed by a brief assessment of the suitability of each approach to the scenarios in which it was tested.

Overall, predictions based upon temporal proximity are found to be more effective than speed-heading predictions, which in turn out are more effective than predictions based upon spatial proximity.

Finally, the chapter concludes with the results of the user evaluation study, which suggests that users of mobile information retrieval tools found the implemented “search ahead” filter useful, are receptive to the idea of other geographic filters, and benefit from the use of personalised geographic information with a spatial and temporal component.

Publication Type: Thesis (Doctoral)
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Departments: School of Science & Technology > Computer Science
School of Science & Technology > School of Science & Technology Doctoral Theses
Doctoral Theses
[thumbnail of Mountain thesis 2005 Vol 2 PDF-A.pdf]
Text - Accepted Version
Download (4MB) | Preview


Add to AnyAdd to TwitterAdd to FacebookAdd to LinkedinAdd to PinterestAdd to Email


Downloads per month over past year

View more statistics

Actions (login required)

Admin Login Admin Login