Geographical aspects of enterobiasis in Estonia.

This paper examines the geographical pattern of the prevalence of enterobiasis among children in Estonian counties, and methods for risk modelling. The methodological questions were as follows: Case-based predictions were more liable to result in over fitting in the case of small training samples than the classification tree models. The spatial autocorrelation of the prevalence was significant up to a distance of 20 km. The main reason for the differences in the predictions might simply be the regional differences in the prevalence.