Classification of EU Countries according to Selected Indicators from the Field of Business Demography Using Self-organising Maps

Abstract

Business Demography monitors the file of active enterprises – those actually doing business regardless of when they obtained the authorisation to operate, closed enterprises – those out of business regardless of when they were actually legally terminated, and enterprises surviving for a given time since their establishment. The result of this monitoring is a group of basic indicators used to characterise the number of newly established, closed and surviving businesses in the EU as well as indicators related to the number of people these enterprises employ.
The main aim of this article is to classify EU countries by selected derived data focused on business demography using a neural network model – Kohonen self-organising maps for the last monitored period. This model involves creating homogeneous groups of countries to be characterised by demographic indicators associated with the birth, survival, and deaths of enterprises and the related employment, all of which are interconnected. This means the classic method of classification cannot be employed. Kohonen maps, however, can be, and may be enabled by various software. We decided to use the tools offered by the statistical analytical system SAS Enterprise Miner (SAS®EM).
An additional objective was to introduce business demography and a description of selected indicators derived in the EU from 2008 to the last reporting period, with a focus on economic development in Slovakia and in the Visegrad Group (V4).