Transforming Manufacturing: A Machine Learning-Driven Approach to Cost Efficiency and Profit Growth

Abstract

Manufacturing industries continually seek innovative strategies to optimize costs and maximize profits. Machine learning (ML) techniques have emerged as a transformative solution, enabling data-driven decision-making and process automation. This study explores the application of ML in manufacturing cost optimization, specifically through two models: Random Forest for predicting Optimized Cost and XGBoost for forecasting Profit After Optimization. Using data from 100 countries, performance metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and R-squared (R²) were used to evaluate the models. The Random Forest model exhibited relatively high error rates, with an MAE of 9846.97, MSE of 142,733,221.95, and a negative R² score of -0.0587, indicating poor model fit and suggesting the need for further refinement. In contrast, the XGBoost model demonstrated superior performance, with an MAE of 309.16, MSE of 161,785.14, and an outstanding R² score of 0.9952, reflecting its strong predictive capabilities. These results underscore the effectiveness of XGBoost in forecasting profit optimization, while highlighting the need for improvements in the Random Forest model for optimized cost prediction. The findings suggest that ML models, particularly XGBoost, have significant potential for cost and profit optimization in manufacturing processes. Future work could focus on enhancing model performance and exploring alternative ML techniques for broader application in manufacturing cost reduction strategies.