SDHO-KGNN: An Effective Knowledge-Enhanced Optimal Graph Neural Network Approach for Fraudulent Call Detection

Abstract

Rapid advancements in mobile communication technologies have led to the progression of telecom scams that not only deplete individual fortunes but also affect social income. Hence, fraudulent call detection gains significance, which not only aims to proactively recognize the frauds, but also alleviate the fraudulent activities to manage external losses. Though the traditional methods, such as rule-based systems and supervised machine learning techniques, actively engage in detecting such fraudulent activities, they fail to adapt to the evolving fraud patterns. Therefore, this research introduces a sheepdog hunt optimization-enabled knowledge-enhanced optimal graph neural network classifier (SDHO-KGNN) approach for detecting fraudulent calls accurately. The effectiveness of the proposed SDHO-KGNN approach is achieved through the combination of the power of graph representation learning with expert insights, which allows the proposed SDHO-KGNN approach to capture complex relationships and patterns within telecom data. Additionally, the integration of the SDHO algorithm enhances model performance by optimizing the discrimination between legitimate and fraudulent calls. Moreover, the SDHO-KGNN classifier captures the intricate call patterns and relationships within dynamic call networks, thereby attaining a better accuracy, precision, and recall of 93.8%, 95.91%, and 95.53% for 90% of the training.