Development of a Hybrid Machine Learning Model for Predictive Performance Optimization in Lean Manufacturing and Industry 4.0

Abstract

The convergence of Machine Learning (ML) and Lean Manufacturing within Industry 4.0 has opened new frontiers for predictive performance optimization in industrial engineering. However, a comprehensive synthesis of existing methodologies and hybrid modeling approaches remains limited in the literature. This study presents a PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses)-based systematic review to identify, evaluate, and synthesize existing research on the integration of machine learning techniques with Lean Manufacturing principles in Industry 4.0 environments. A structured database search was conducted across Scopus, Web of Science, and Google Scholar, yielding an initial pool of 320 articles, of which 47 studies were selected following strict inclusion and exclusion criteria. The review systematically examines supervised, unsupervised, and hybrid ML algorithms applied to key Lean metrics including waste reduction, cycle time, defect prediction, and process efficiency. Findings reveal that hybrid ML models combining algorithms such as Random Forest, XGBoost, and Neural Networks demonstrate superior predictive accuracy compared to single-algorithm approaches. Furthermore, the study identifies critical research gaps in real-time data integration, model interpretability, and scalability across manufacturing sectors. Based on the synthesized evidence, a conceptual hybrid ML framework is proposed to guide future model development for predictive Lean performance optimization. This review contributes a structured foundation for researchers and practitioners seeking to advance data-driven decision-making in smart manufacturing systems.