Spatio-temporal outlier detection based on context: a summary of results

Authors:
Zhanquan Wang;Chao Duan;Defeng Chen
Affiliations:
Department of Computer Science and Engineering, East China University of Science and Technology, Shanghai, China;Department of Computer Science and Engineering, East China University of Science and Technology, Shanghai, China;Department of Computer Science and Engineering, East China University of Science and Technology, Shanghai, China
Venue:
AICI'11 Proceedings of the Third international conference on Artificial intelligence and computational intelligence - Volume Part I
Year:
2011

Citing 5
Cited 0

Data mining: concepts and techniques

Data mining: concepts and techniques
Outlier Detection Using Random Walks

ICTAI '06 Proceedings of the 18th IEEE International Conference on Tools with Artificial Intelligence
Conditional Anomaly Detection

IEEE Transactions on Knowledge and Data Engineering
Isolation Forest

ICDM '08 Proceedings of the 2008 Eighth IEEE International Conference on Data Mining
Discovering Contexts and Contextual Outliers Using Random Walks in Graphs

ICDM '09 Proceedings of the 2009 Ninth IEEE International Conference on Data Mining

Quantified Score

Hi-index	0.00

Visualization

Abstract

Spatio-temporal outlier detection plays an important role in some applications fields such as geological disaster monitoring, geophysical exploration, public safety and health etc. For the current lack of contextual outlier detection for spatio-temporal dataset, spatio-temporal outlier detection based on context is proposed. The pattern is to discover anomalous behavior without contextual information in space and time, and produced by using a graph based random walk model and composite interest measures. Our approach has many advantages including producing contextual spatio-temporal outlier, and fast algorithms. The algorithms of context-based spatio-temporal outlier detection and improved method are proposed. The effectiveness of our methods is justified by empirical results on real data sets. It shows that the algorithms are effective and validate.