Optimizing Fraud Detection in Digital Banking with a Novel Hybrid Framework

Digital banking is becoming more popular as more people use smartphones, laptops, and other devices, making banking and financial transactions faster and easier.However, this increase in use also attracts fraudsters who exploit weaknesses in these systems.As a result, banking fraud is rising, particularly in developing countries where the quick adoption of technology has outpaced the development of strong security measures.In these regions, weak security has caused potential customers to hesitate, raising concerns about the safety of financial data.Fraud, which is not new, continues to cause significant global financial losses, costing billions of dollars annually.The issue is worsened by the fact that most transaction data is marked as legitimate, making it difficult to detect fraud.To combat this, researchers have developed various fraud detection methods using machine learning, such as LGBM, random forests, SVM, deep learning, neural networks, XGBoost, and logistic regression.These models are tested on transaction data to identify suspicious activity.However, when trained on imbalanced data, these models tend to favor legitimate transactions, making them less effective at detecting fraud.This study addresses this by using machine learning classifiers to detect fraud in banking payments and transfers.The data used reflects real-time transactions simulating common fraud schemes.In this research, we extend our previous work on 'Machine Learning Based Predictive Modelling for Fraud Detection in Digital Banking'.This research builds on previous work in fraud detection and introduces a new approach combining Hidden Markov Models (HMM) with ensemble machine learning methods, hyperparameter tuning, and the SMOTE technique to improve fraud detection accuracy.Our results show that this approach outperforms traditional models, enhancing both detection accuracy and efficiency.The findings provide insights into improving fraud detection in digital banking services.