APPLYING LSTM DEEP LEARNING MODEL TO PREDICT CORPORATE CAPITAL STRUCTURE: A DATA-DRIVEN APPROACH

Abstract

Capital structure plays a crucial role in corporate finance, influencing investment decisions and financial performance. Traditional econometric models have been widely used to predict capital structure; however, recent advancements in deep learning offer promising alternatives. This study explores the application of Long Short-Term Memory (LSTM) models for capital structure prediction, comparing their performance against conventional machine learning approaches. Using financial data from major corporations, we analyze the predictive capabilities of LSTM architectures. Our findings indicate that LSTM models exhibit superior accuracy in capturing complex patterns within financial datasets, demonstrating their potential as powerful tools for capital structure forecasting.