APPLYING MACHINE LEARNING IN PRODUCTION: A SUBJECT FIELD REVIEW

The past decade has shown a continuing interest in machine learning and its potential for use in production. In response to the rapidly increasing number of publications, a subject field review aimed to summarize the existing literature on the application of machine learning in different stages of production. This review explores the possible applications of various machine learning methods currently in existence at different stages of production. The searches were conducted in Scopus electronic database from 2019 to 2021, this period was not covered in previous reviews. Also, a distinctive feature of this review is the classification of machine learning applications by stages of the production process and the accounting of studies by country. A total of 2121 papers were identified, 35 articles were selected based on the predetermined eligibility criteria and included in this review. Most of the papers described the application of machine learning during the production planning stage. In addition, it turned out that the most used is supervised learning. Most of the publications on the application of machine learning methods in production were created by researchers from China. The papers before 2019 indicated that the most widespread and most successfully applied is the Bayesian networks, while it is currently considered the most widespread is supervised learning. To identify patterns of machine learning applications at specific stages of production, it would be desirable to conduct more research in other countries in order to reduce the level of influence of research from China on general trends. The review can be used by the professionals interested in machine learning processes.