A Shape‐Based Local Spatial Association Measure (LISShA): A Case Study in Maritime Anomaly Detection. Issue 4 (19th November 2018)

• Record Type: Journal Article
• Authors: Roberts, Steven Andrew
• Is Part Of: Geographical analysis. Volume 51:Issue 4(2019)
• Page Start: 403
• Page End: 425
• Publication Date: 2018-11-19
• Subjects: Geography -- Methodology -- Periodicals; Electronic journals; 910.00182
• DOI: 10.1111/gean.12178
• Languages: English
• ISSNs: 0016-7363

Abstract

Abstract : The explicit consideration of the shape of geographic features has been largely ignored in existing spatial association measures. The primary contribution of this work is the development of a new local spatial association measure—a Local Indicator of Spatial and Shape Association (LISShA). The LISShA measure is modeled after local Geary's Spatial Autocorrelation measure with distance between shapes, calculated using the Small–Le metric, replacing difference between attribute values and the spatial neighborhood defined by Fréchet distance. We provide some explanation of these metrics and show, in detail, how the LISShA and proposed moments are calculated in a one‐dimensional context in a case study of maritime anomaly detection.