Evaluation of Hotspot Detection Method based on Echelon Structure

The study for finding a hotspot, such as disease clustering or hazard map is one of important issue. A spatial scan statistics based on the likelihood ratio associated with the number of events inside and outside a circular scanning window has been widely used as a hotspot detection method. However, it is noted that a non-circular shaped hotspot, such as the shape made by a river or a road cannot be detected. We have proposed a technique using an echelon analysis as a non-circular shaped hotspot detection. The echelon analysis is a useful technique for systematically and objectively investigating the phase-structure of spatial regional data. In this paper, we evaluate the validity of echelon’s hotspot detection by comparing with the result of all possible scanning method which certainly detect the hotspot with the highest likelihood ratio.