Predicting Firm Value Using Ensemble and Nonlinear Machine Learning Models: Evidence from Financial and Value-Based Metrics

Abstract

This study aims to predict firm value using financial indicators, economic profit, and intangible capital through machine learning approaches. The independent variables include precautionary cash, leverage, asset utilization, short-term liquidity ratio, economic profit, and intangible capital, while firm value is measured using Price-to-Book Value (PBV). This research employs several machine learning models, including Linear Regression, Decision Tree, Random Forest, Gradient Boosting, Neural Network, and Support Vector Machine. Model performance is evaluated using Mean Squared Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination (R²). The results show that the Random Forest model provides the best predictive performance, explaining approximately 90% of the variation in firm value. Asset utilization emerges as the most influential variable, followed by short-term liquidity ratio and economic profit. Meanwhile, leverage and precautionary cash show relatively smaller contributions to firm value prediction. These findings indicate that firm value is primarily influenced by operational efficiency, liquidity performance, and value creation capability. The study demonstrates that machine learning methods provide a comprehensive and effective approach to predicting firm value using financial and value-based performance indicators.