Journal of Economics and Management Volume 21, No. 2 September, 2025

Predicting and Analyzing the Intangible Assets Value with Machine Learning Techniques

Yu-Hsin Lu

Department of Accounting, Feng Chia University, Taichung, Taiwan

Yu-Feng Hsu

Department of Accounting and Information Technology, National Chung Cheng University, Chia-yi, Taiwan

Abstract

With the shift to a knowledge-based economy, the importance of intangible assets to management has increased. Therefore, evaluating the real value of intangible assets has become a critical issue. This study develops and compares two models based on machine learning techniques: single classifier models and ensemble learning methods. The results show that the ensemble learning methods do not perform better than the single best classifiers. On the other hand, a well-performing single classifier model produces the optimal intangible asset value prediction model. Otherwise, this study shows and explains essential features in different periods. Finally, moving window-based and fixed window-based evaluation methods are adopted to verify the prediction algorithms, since the dynamic economy and business environment changes over time may affect the data. The results show that the 1Y Random Tree method provides the best results. These prediction models can provide valuable information to management and shareholders to evaluate the status of companies and make more effective decisions.

Keywords：Intangible Assets Value, Artificial Neural Networks, Linear Regression, Support Vector Regression, Ensemble Learning

JEL Classifications：M41, M48

BACK