Cluster-based congestion outlier detection method on trajectory data
FSKD'09 Proceedings of the 6th international conference on Fuzzy systems and knowledge discovery - Volume 5
MEC --Monitoring Clusters' Transitions
Proceedings of the 2010 conference on STAIRS 2010: Proceedings of the Fifth Starting AI Researchers' Symposium
Bipartite graphs for monitoring clusters transitions
IDA'10 Proceedings of the 9th international conference on Advances in Intelligent Data Analysis
ciForager: Incrementally discovering regions of correlated change in evolving graphs
ACM Transactions on Knowledge Discovery from Data (TKDD)
| Hi-index | 0.00 |
Moving objects are becoming increasingly attractive to the data mining community due to continuous advances in technologies like GPS, mobile computers, and wireless communication devices. Mining spatio-temporal data can benefit many different functions: marketing team managers for identifying the right customers at the right time, cellular companies for optimizing the resources allocation, web site administrators for data allocation matters, animal migration researchers for understanding migration patterns, and meteorology experts for weather forecasting. In this research we use a compact representation of a mobile trajectory and define a new similarity measure between trajectories. We also propose an incremental clustering algorithm for finding evolving groups of similar mobile objects in spatio-temporal data. The algorithm is evaluated empirically by the quality of object clusters (using Dunn and Rand indexes), memory space efficiency, execution times, and scalability (run time vs. number of objects).