Predictive Analytics and Financial Forecasting with AI

This paper explores the use of advanced machine learning techniques for financial forecasting, focusing on stock price prediction. Traditional econometric models often struggle with the nonlinear dynamics of financial markets. This study employs Linear Regression, Random Forest, Gradient Boosting, and Feedforward Neural Networks (FFNNs) to improve predictive accuracy. Data covers financial metrics of 207 companies across 12 industries for 2022 and 2023. Key variables include estimated earnings per share, long-term debt, EBITDA, interest expense, return on equity, quick ratio, price-earnings ratio, and stock price volatility. The models are evaluated using Mean Squared Error (MSE) and Mean Absolute Percentage Error (MAPE). Results show that Random Forest and Gradient Boosting outperform Linear Regression. FFNNs, particularly with Dropout layers, achieve the lowest MAPE. This study highlights the effectiveness of machine learning in financial forecasting, providing insights into model selection based on data characteristics and predictive goals.