The Impact of the Training Set Size on the Classification of Real Estate with an Increased Fiscal Burden

Abstract

Abstract The introduction of an ad valorem tax can lead to an increase in the tax burden on real estate. There are concerns that this increase will be large and widespread. Before undertaking any actual actions related to the real estate tax reform, pilot studies and statistical analyses need to be conducted in order to verify the validity of those concerns and other aspects regarding the replacement of a real estate tax, agricultural tax and forest tax with an ad valorem tax. The article presents results of research on the effectiveness of the classification of real estate into a group at risk of an increase of tax burden with the use of the k-nearest neighbors method. The main focus was to determine the size of a real estate set (training data set) on the basis of which classification is conducted, as well as on the efficiency of that classification, depending on the size of such data set.