Development of the algorithm for classifying industries according to the type of intra-factory cooperation of main and auxiliary processes using machine learning

Abstract

The task of rational organization of auxiliary processes at the enterprise is to reduce their cost by deep integration into the main production process. The purpose of the article is to develop a classification analysis algorithm for assessing the dependencies between the main and auxiliary units and the typology of production processes according to the level of intra-factory cooperation. As a method for determining the type of production, the Random Forest machine learning method using the bagging machine learning meta-algorithm is proposed. Parameters have been developed that describe the costs of auxiliary operations, the costs of repair facilities and equipment maintenance, the level of technical efficiency of production. Approbation of the algorithm on the example of chemical enterprises made it possible to distinguish three types of production according to the nature of intra-plant cooperation of processes according to the most informative parameters. To assess the usefulness and performance of the models, cumulative lift diagrams were constructed, where the most productive type was determined with an average level of intra-factory cooperation. The results are the primary diagnostics of the organization of auxiliary facilities, decision-making on the reengineering of processes in order to strengthen intra-factory cooperation and reduce costs.