Long-Short-Term Memory Neural Networks and the Evolution of Financial Forecasting, Fraud Detection, and Audit Innovation: A Narrative Review Study

This narrative review explores the application of Long Short-Term Memory neural networks in financial forecasting, fraud detection, and auditing. It addresses the limitations of traditional methods in capturing complex temporal relationships within financial data. The review synthesizes findings from recent studies, highlighting advancements in LSTM architectures, such as Bidirectional and Convolutional LSTMs, which significantly improve predictive accuracy. Additionally, it discusses the effectiveness of LSTM models in detecting fraudulent activities, achieving higher precision and recall compared to conventional techniques. Furthermore, the integration of LSTMs with blockchain technology and explainable AI has enhanced auditing practices, facilitating real-time analysis and improving reliability in financial reporting. Overall, this review underscores the transformative potential of LSTM networks in modern finance.

کلیدواژه ها:

Long-Short-Term Memory ، Financial Forecasting ، Fraud Detection ، Audit Innovation