Determinants of crime rate in Poland. Analysis using regression trees

The aim of this article is to identify relationships between crime rate and some socio-economic, demographic and environmental factors in the poviats of Poland. There were analysed cross-sectional data using regression tree. The following factors were found to significantly explain the intensity of crime rate: urbanisation, percentage of single-person households, provided accommodation per 1000 population, divorce’s coefficient and the intensity of crime in the neighboring areas. Then the random forest was used to improve prediction’s accuracy and generate rank of variable importance.