Data mining for visual exploration and detection of ecosystem disturbances

This paper presents a case study on the application of data mining to the problem of detecting ecosystem disturbances from vegetation cover data obtained from satellite observations. We describe two anomaly detection approaches--moving average and random walk--for detecting such events. We also illustrate how clustering can be used to locate similar incidents of disturbance events. Finally, we present a clustering-based framework to aid the visual exploration of ecosystem disturbances from high resolution data.

• Type: Conference
• Year: 2008
• Where: GIS
• Authors: Haibin Cheng, Pang-Ning Tan, Christopher Potter, Steven A. Klooster
• Keywords: Database, Detection Approaches--moving Average, Ecosystem Disturbances, GIS 2008, Vegetation Cover Data