Spatial model selection based on hybrid performance measure of linear classifier

Abstract

Assuming that spatial data is generated by Gaussian random field (GRF), the problem of classifying its observation into one of two populations is considered. Populations are specified by the common regressors but different regression parameters. Authors concern with classification procedures associated with Bayes discriminant function (BDF) and its sample version (SDF). The average of plugin and apparent correct classification rates is considered as performance measure of classifier based on SDF. Various types of spatial data models for invasive species (zebra mussels) distributed in the Curonian Lagoon are considered and ranked by the defined criterion. Advanced models are proposed to the mapping of presence and absence of zebra mussels in the Curonian Lagoon.