Evaluation of Industry 4.0 familiarity at SMEs in Central-Eastern Europe using Machine Learning Algorithms

Abstract

Familiarity with Industry 4.0 (I4.0) and its components is inevitable in business development, since it supports digitalization among the various sectors of businesses SMEs in Central-Eastern Europe, namely in the V4 countries, Serbia and Bulgaria have been surveyed; how familiar they are with I4.0 and its components. This paper presents how machine learning (ML) methods, namely supervised algorithms help to determine which I4.0 components contribute positively to SMEs’ familiarity with I4.0 and which of them contradict to its familiarity. The ML algorithm Vector Support Machine (VSM) and Neural Network (NN) proved to be the most accurate methods to predict the familiarity with I4.0 based on its components while Decision Tree (DT) gave the highest precision and specificity rates, therefore, the predictions of these three methods are compared. The results imply that Cloud Computing Services, Big Data analysis, VR and 3D Printing and Robotics do not contribute to the familiarity with I4.0 as much as expected. SMEs need further information on I4.0 developments, especially on its components’ beneficial impact on business performance and operations.