Extracting Places from Traces of Locations Algorithms featured by: Jong Hee Kang William Welbourne Denjamin Stewart Gaetano Dorriello Present by: Yuan Liu
Introduction What are these algorithms for? Extracting locations from cluster points What are these algorithms trying to achieve? Significantly reduce the server load Maintain data accuracy Translate coordinates to locations
Recall of My Senior Design problem
Recall of My Senior Design problem (continue) What did our customer wants: Plug collected GPS points into Google Map Can group these points manually or automatically Customize these grouped points (name, etc) Our Approach: Use Area grouping (draws a circle) Use a automatic filter
Article proposed Algorithms Time based Clustering: clustering locations along the time axis New location is compared with previous locations If the new location is moving away starts a new cluster Then ignore the clusters with short time duration
Article proposed Algorithms (continue)
Article proposed Algorithms (continue) Frequency clustering: Time based Clustering algorithm only consists data with longer time duration People may visit important place frequently, but not for a long time (such as ATM, mailbox) Need two threshold value: one determines duration, the other determines the frequency of visiting. May yield false information (stop at traffic light)
How would these algorithms may help my team Wheelchair people usually does not move around in a big radius, and they tend to stay at one place for a long period of time, therefore the Time clustering algorithms applies the best. The server we were using is quite weak, and we didn’t assume everyone has broadband. Such algorithms will greatly reduce the calculation time.
Questions?