SPATIAL CROSS-VALIDATION AND BOOTSTRAP FOR THE ASSESSMENT OF PREDICTION RULES IN REMOTE SENSING: THE R PACKAGE SPERROREST

Novel computational and statistical prediction methods such as the support vector machine are becoming increasingly popular in remote-sensing applications and need to be compared to more traditional approaches like maximum-likelihood classification. However, the accuracy assessment of such predictive models in a spatial context needs to account for the presence of spatial autocorrelation in geospatial data by using spatial cross-validation and bootstrap strategies instead of their now more widely used non-spatial equivalent. These spatial resampling-based estimation procedures were therefore implemented in a new package 'sperrorest' for the open-source statistical data analysis software R. This package is introduced using the example of the detection of rock-glacier flow structures from IKONOS-derived Gabor texture features and terrain attribute data.