AI in Financial Management in Saudi Arabia

Abstract

Abstract—Financial management is vital for organizational stability and strategic decisions, especially in Saudi Arabia’s dynamic, data-driven economic environment. Conventional financial prediction models are limited by market volatility, data inconsistency, and complex interdependencies among financial indicators, resulting in lower accuracy. This research overcomes these challenges using an artificial intelligence (AI)-driven framework that integrates optimization techniques, attention-based feature learning, and deep neural modeling. The financial management datasets were collected from Saudi Arabian financial systems. Then, Min–Max normalization and missing-value handling were applied to standardize data and improve learning stability. Discrete Wavelet Transform (DWT) was utilized to extract meaningful time-frequency financial features, enhancing signal representation and reducing noise. Mayfly Optimized–Driven Attention Deep Neural Network (MFO-Att-DNeuroNet) was proposed to enhance financial prediction accuracy and model robustness in financial management applications. The Att-DNeuroNet selectively emphasized the most informative financial features while modeling complex nonlinear relationships within the data. MFO was employed to optimally tune model parameters, ensuring faster convergence and improved predictive performance. Experimental evaluations demonstrated that the MFO-Att-DNeuroNet achieved superior prediction accuracy (0.97) when using Python 3.10. The MFO-Att-DNeuroNet provided an effective and scalable solution for intelligent financial management, offering significant potential to support accurate forecasting and strategic decision-making in Saudi Arabian organizations.