Browse Prior Art Database

A System & Method for Predicting Future Spatial and Temporal Occurrence of Users based on their Past Spatio-Temporal Behavior

IP.com Disclosure Number: IPCOM000245213D
Publication Date: 2016-Feb-18
Document File: 4 page(s) / 171K

Publishing Venue

The IP.com Prior Art Database

Abstract

Given a compact representation (signature) that captures hangouts and movement patterns of each user distinctly, proposed system can predict the possible location(s) at which the user could be present at a given time, as well as the possible time in future at which the user could visit a given location(s).

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 43% of the total text.

Page 01 of 4

A System & Method for Predicting Future Spatial and Temporal Occurrence of Users based on their Past Spatio-Temporal Behavior

Problem: Existing techniques for processing spatio-temporal data do not provide a technique to

predict location of a user at a given time, or future timestamp(s) when a user would visit a query location, given the spatio-temporal signatures, or to visualize these predictions graphically.

Proposal: Given a Spatio-temporal signature of a user, based on the nature of the query, following prediction is performed


- Spatial Prediction Query - For the location(s) provided, access the corresponding locations in the ST-signature of the user, and perform a regression (or any valid) time-series analysis process to determine the next occurrence of the user at this location(s)


- Temporal Prediction Query - Perform a (efficient) scan on the ST-Signature and based on the periodicity time interval to which the specified input time maps, determine the set of possible location(s) at which the user could exist

Known Solutions & Their Drawbacks: Existing works in this domain can be classified and differentiated from our proposed system as follows:

Spatio-Temporal Prediction For Trajectories (With History)


- WhereNext: A Location predictor on trajectory pattern mining (http:// dl.acm.org/citation.cfm?id=1557091)

- Semantic trajectory mining for location prediction (http:// dl.acm.org/citation.cfm?id=2093980)


- Primary focus is on finding interesting global patterns and answering queries of the type "if user is at location X at time T1, where will he be at time T2".


- Ignores user-centric historical movements and requires a current user location to

predict a future location.

- Does not address queries of the type - "where will user visit location X next?" or "independent of current location, where will user be at a future point of time?"

Real-time Prediction For Trajectories (No History)


- A Predictive Location Model for Location-Based Services ( http://dl.acm.org/citation.cfm?id=956693)


- Accuracy and Resource Consumption in Tracking and Location Prediction ( http://www.cs.uic.edu/Ċµolfson/mobile_ps/sstd03.pdf)


- Given current trajectory, an immediate future location is predicted using road-network and/or Euclidean distance computations by taking speed and direction in account


- However, fails to address the kind of queries we are interested in (mentioned above) since it does not involve understanding the historical movement patterns and hangouts of users

Summary of Method:

1


Page 02 of 4

Input


- For each user, his/her spatio-temporal data (GPS traces, CDR records, etc.) collected over time that represents his/her behavior or signature in terms of movement patterns and hangouts

Output


- Given a specific time in future, the possible location(s) at which the user could be

  present
- Given a specific location(s), the possible time in future at which the user could

visit this location(s)

Novelty


- Spatial Prediction using Periodicit...