AI in Equity Valuation Models: Enhancing Valuation Accuracy Using Machine Learning Approaches
0 viewsAbstract
This paper will discuss whether artificial intelligence-based valuation models are better predictors of the market capitalization of a listed company (when compared to a traditional linear regression model). A list of 20 large publicly traded firms in the U.S. was put together based on SEC 10-K annual reports of the 2023 fiscal year, across six industries such as technology, healthcare, consumer goods, financials, energy, and the industrial sectors. Three important equity valuation financial drivers were identified and calculated using official filings, which are the return on equity (ROE), revenue growth, and size of the firm as the logarithm of total assets. The dependent variable was market capitalization as of December 2023 and was taken from Yahoo Finance. An estimated traditional linear regression model was used as a benchmark and then an AI-based regression model that considers non-linear relationships between the financial variables was estimated. R- squared, Root Mean Squared Error (RMSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) were used to measure model performance. Linear regression had an R- squared of 0.512 and MAPE of 84.3 and AI-based model had an R-squared of 0.847 and MAPE of 41.6%. These findings verify that machine learning-based valuation models have a significant increase in predictive accuracy, especially in high-capitalization and high-growth companies where non-linear market dynamics are the most significant.
Keywords
Artificial Intelligence, Equity Valuation, Financial Modeling, Linear Regression, Machine Learning, Market Capitalization, Return On Equity, Revenue Growth