Machine Learning-Based Budget Prediction for Quantitative Decision-Making

Abstract

Accurate budget prediction is a fundamental challenge for quantitative decision-making in both public and private sectors. Traditional statistical approaches often fail to capture the complex, nonlinear relationships that characterize modern financial and economic data. This chapter investigates the effectiveness of machine learning-based regression models for budget prediction using three heterogeneous real-world datasets related to construction costs and film production budgets. Linear Regression, Random Forest, and Gradient Boosting models are implemented and evaluated using standard performance metrics, including MAE, RMSE, and the coefficient of determination (R2). The experimental results demonstrate that ensemble-based models significantly outperform linear methods across all datasets, with Gradient Boosting providing the highest predictive accuracy and robustness. The findings confirm the nonlinear nature of budget estimation problems and highlight the strong potential of machine learning as a numerical tool for supporting quantitative financial decision-making.