Machine Learning Models for Predictive Optimization in Data-Driven Engineering Systems

Abstract

This work introduces a predictive optimization framework with machine learning to improve the performance of data-driven engineering systems. Development and testing of several models, including Linear Regression, Support Vector Machine (SVM), Random Forest, and Gradient Boosting, were developed and tested using a structured dataset containing key variables, such as system load, energy consumption, efficiency, and risk of failure. The models were evaluated based on the accuracy, RMSE, and R 2 values, with the highest accuracy being 92.4% with the lowest prediction error of Random Forest and then Gradient Boosting. The analysis of the importance of the features showed that the most influential are the system load and the energy consumption, which have the greatest effect on the optimization results. Results of cross-validation validated the robustness and reliability of ensemble models on various data subsets. The results have shown that machine learning methods are powerful predictors and decision-makers in the field of engineering. The proposed framework provides a scalable predictive maintenance, resource optimization, and intelligent system management solution in real-world engineering settings.