Artificial Intelligence for Accounting: How can Machine Learning Models predict firm’s profitability?

The accurate prediction of a company’s future earnings could represent a significant milestone in finance, particularly in accounting research and investment practice. The possible range of applications of such an idea could vary from aiding investment decision making to evaluating corporate performance (Lev and Gu, 2016). Investors could utilize predictive models to identify undervalued stocks and optimize their portfolios, while corporate managers might leverage these insights to enhance strategic planning and resource allocation. Furthermore, regulators and policymakers could use earnings predictions to monitor market stability and inform economic policy. However, a company’s future performance is influenced by numerous concurrent and complex factors, including macroeconomic conditions, industry trends, management decisions, and unforeseen events such as technological disruptions or regulatory changes; and past literature does not provide clear guidance on effective proxies for predicting future earnings (Lev and Gu, 2016; Monahan, 2018), often due to the diverse methodologies and varied contexts in which studies are conducted. Despite advancements in financial modelling and data analytics, the inherent uncertainty and multifaceted nature of business environments pose significant challenges to developing universally applicable predictive frameworks. Therefore, ongoing research is essential to refine these models and incorporate more dynamic and comprehensive variables that can better capture the complexities of real-world economic scenarios.