Abstract
In this study, a novel hybrid analytical framework between the regression tree (RT) from machine learning and classical econometrics analysis is presented to understand the relationship between financial parameters and enterprise value (EV). Five predictors, which are: return on equity (ROE), debt-to-asset ratio (DAR), institutional ownership (IO), firm size, and firm age, were examined as independent variables. In particular, the algorithm was applied to split a dataset of 52 Indonesian consumer sector firms, dated from 2018 up to 2023, where then a linear regression model was assigned to each cluster of data. Based on the conducted numerical simulations, it was concluded that ROE and firm size had a consistently positive effect on the EV across all clusters. Meanwhile, the effect of IO, DAR, and age varied on each node. These findings suggest that the relation between financial parameters and firm value (FV) is not uniform and can be interpreted better by considering multi-segment data. This method serves as a new data-driven methodology to the traditional panel analysis, which is complex and requires significant knowledge in analytical statistics.
Keywords:
- Keyword: Enterprise value
- Keyword: Firm value
- Keyword: firm size
- Keyword: Machine Learning
- Keyword: Regression Tree
How to Cite:
Santoso, H. & Tjen, J., (2026) “Regression Tree-Based Segmentation of Enterprise Value: Bridging Machine Learning and Classical Financial Analysis”, Journal of Financial and Economic Dynamics (JFED) .