Data mining methodology for finding ergonomic reserves to increase management efficiency. Method for preventing errors by ERP system operators

Abstract

The article considers the task of increasing the efficiency of automated information processing for enterprise resource planning (ERP) systems. It is shown that in the era of digital transformation, ERPs are becoming critically important information structures that support the automation and coordination of business processes. It is demonstrated that most approaches to ERP reliability analysis, which focused, as a rule, on technical aspects such as data security, software stability or infrastructure stability, do not provide effective means of finding ways to improve ERP efficiency. It is shown that the human factor remains the main source of operational risks. Poor interface ergonomics, excessive navigation depth, lack of contextual guidance or cognitive overload often lead to errors, reduced efficiency and disruption of data integrity. In connection with the growing threats of neglecting the human factor, the article considers the task of identifying hidden patterns between user interaction, interface characteristics and the probability of errors. The methodology of data mining is used, in particular the principle of associative rules. It is shown that, unlike traditional statistical approaches, association rules allow to detect non-obvious dependencies in user behavior, forming the basis for adaptive, proactive decision support. An experimental ERP environment is proposed for collecting real interaction logs, user profiles, and error reports. The generation of over 500 association rules is demonstrated, of which 64 provided high stability and repeatability, which allowed the development of a decision support module integrated into the ERP interface. Empirical results show significant improvements: a 24% reduction in critical errors, a 12% reduction in task execution time, and a 23% reduction in cognitive load (NASA-TLX). The results confirm the potential of association rule analysis as a methodological basis for designing adaptive ERP systems, with an emphasis not only on technical but also on ergonomic and cognitive aspects of human-machine interaction. The results have been used in ergonomic ERP design in industry, agricultural production, and e-learning. It is proposed to use the results in the international scientific project "International Center for the Support of Ergonomic Design and Research" within the framework of the International Ergonomics Association.