| Summary: | The current COVID-19 pandemic has shown how important digitalization is to the world economies and how networks
and connectivity, artificial intelligence, as well as basic and advanced digital skills sustain economies and societies. Digitalization involves an increase in the use of information and communication technologies (ICT) in all areas of the economy and all domains of a functioning society. This study was based on the Digital Economy and Society Index (DESI) dimensions: Connectivity; Human capital; Use of internet service; Integration of digital technology; Digital public services for 2020. The data included in the index were mostly collected from the relevant authorities of the Member States by the European Commission (Directorate-General for Communications Networks, Content and
Technology as well as Eurostat). The Kohonen neural network technique, which uses an iterative algorithm, was used to group countries according to the DESI dimension. It is a method that enables grouping of objects into sets due to the similarity of the data they contain and characterizes them. In this study, the EU countries were divided into groups according to the similarity of DESI levels. For the set of 28 countries and 5 DESI dimensions the appropriate structure of the Kohonen network and its parameters were selected, in order to determine the most satisfactory results. The following were the groups of countries classified: Class ID1: Denmark, Finland, Netherlands, Sweden, Class ID2:
Belgium, Estonia, Ireland, Malta, United Kingdom, Class ID3: Austria, Germany, Spain, France, Lithuania, Luxembourg, Latvia Class ID4: Czechia, Hungary, Italy, Poland, Portugal, Slovenia, Class ID5: Bulgaria, Cyprus, Greece, Croatia, Romania, Slovakia. The distribution of the separate classes of the state indicates that we can treat them due to the specificity of their features. In the ID 1 and ID2 classes the considered features assume very high numerical values. ID class 3 is characterized by an average level of DESI parameters. Low values of the considered features are characteristic of the ID4 and ID 5 classes. The classification result is helpful in distinguishing the current level of digitalization of European Union countries.
|
|---|