A Robust SVR-Based Model for Predicting Tooling Costs in Manufacturing

In the last decade, intensified competition, global risks, and reliance on competitors’ pricing strategies in the truck and bus industry have compelled companies to monitor and track their costs with increased precision. Stringent cost monitoring, calculation, and budget management are essential to ensure sustainable growth and long-term profitability, thereby elevating the significance of the cost management department. Additionally, companies strive to shorten the time-to-market to outpace competitors and maximize early market opportunities. The rapidly changing environment necessitates that cost management adopt shortened, iterative, and renovated approaches. In recent years, various ML-based models have been developed to address these challenges. Among other machine learning models, support vector machine (SVM) models come forward for parametric cost estimation because of their robustness to noise and outliers, efficacy with non-linear relationships, and good success rate with small datasets. Rotational molding has become an increasingly preferred manufacturing method due to the growing demand for low-volume plastic parts. This paper focuses primarily on the development of an improved SVM model with various ensemble methods that capture the cost estimation relationship between the characteristics of rotational molded parts and rotational molds. ... mehrThe proposed ensemble model achieved a mean absolute percentage error (MAPE) of 31.3% (95% CI: 28.2–34.4), indicating an average deviation of approximately ±31% from actual tooling costs. Hence, the developed model was offered for the use of cost engineers via the development of a user-friendly graphical user interface (GUI), which allows them to benefit from the model in the early development phases.