Research on Application of Whale Optimization Algorithm in Financial Payment Fraud Detection

Due to the rise in fraudulent activity in financial payment systems, accurate and efficient methods of fraud detection have become paramount for financial security. To address the limitations of traditional machine learning methods on complex and high dimensional financial data, this research presents a Whale Optimization Algorithm (WOA) based financial payment fraud detection model to enhance the detection performance. On the one hand, in order to solve the convergence speed and local optimality of WOA, this paper proposes an adaptive weight adjustment strategy, which can dynamically adjust the movement strategy of individual whales according to the feedback in the search process, so as to avoid falling into local optimal and accelerate global search. Secondly, in order to enhance the stability of the algorithm, the simulated annealing mechanism is integrated with WOA to further optimize the global search ability. Finally, the adaptive feature selection method is integrated into the model, which is used to optimize the feature subset through WOA, retain the features highly related to fraudulent behavior, and reduce the interference and influence of redundant information on the model. Experimental results demonstrate that the improved WOA algorithm achieves higher detection rate and lower false positive rate compared to the traditional WOA and other commonly used algorithm in multiple real financial payment dataset fraud detection tasks.