Neill’s fast subset scan2 detects significant spatial patterns of disease by efficiently maximizing a log-likelihood ratio statistic over subsets of locations, but may result in patterns that are not spatially compact. The penalized fast subset scan (PFSS)3 provides a flexible framework for adding soft constraints to the fast subset scan, rewarding or penalizing inclusion of individual points into a cluster with additive point-specific penalty terms. We propose the support vector subset scan (SVSS), a novel method that iteratively assigns penalties according to distance from the separating hyperplane learned by a kernel support vector machine (SVM). SVSS efficiently detects disease clusters that are geometrically compact and irregular.
Objective
We present the support vector subset scan (SVSS), a new method for detecting localized and irregularly shaped patterns in spatial data. SVSS integrates the penalized fast subset scan3 with a kernel support vector machine classifier to accurately detect disease clusters that are compact and irregular in shape.