Enhancing FinTech-Driven Digital Transformation in Financial Institutions via Attention-Based Recurrent Neural Networks

<jats:p>FinTech involves the use of current technologies in the financial sector. Recent technologies transform financial services by enhancing transaction security, lowering transaction costs, and improving accessibility. The quick growth of FinTech has altered traditional financial organization processes, accelerating global economic development. The financial industry is rapidly adopting Artificial Intelligence (AI) because of its potential for high cost savings in financial services. Saudi Arabia is developing as an aggressive market in this industry. This paper presents an Attention-based Recurrent Neural Network for Fintech-Driven Digital Transformation in Financial Institutions (ARNN-FDTFI) approach as a prediction framework to forecast financial outcomes in Saudi financial institutions. In the proposed ARNN-FDTFI approach, a preprocessing step is initially performed to clean the data, encode features, detect outliers, and normalize the data to ensure quality. A modified ReliefF-based feature selection strategy is employed to select the most relevant features and eliminate redundancy. For stock exchange prediction, an attention-enhanced Bidirectional Long Short-Term Memory model (Bi-LSTM) is employed to capture temporal dependencies in financial time-series data. The RMSProp optimizer is employed during training to enhance model stability and ensure faster convergence. A comprehensive experimental analysis on the Saudi Stock Exchange (Tadawul) dataset demonstrated the improved performance of the ARNN-FDTFI model compared to existing techniques, with an R2 score of 0.8534.</jats:p>