The Multi-Verse Optimization Algorithm using a Fuzzy Clustering Technique to Effectively Detect Frauds in Automobile

The growing use of the internet, especially in the automotive industry, has resulted in the development of several online services. Globally, the exponential expansion of financial fraud has resulted in huge financial losses for organisations in a variety of industries. This research develops an innovative approach that utilizes a capability of the Multi-Verse Optimization Algorithm (MVOA) to improve the performance in fraud detection through fuzzy clustering method. The contributions of MVOA are in the enhancement of model adaptability to ever-changing fraud patterns, the adept handling of imbalanced data through diverse exploration, and the resolution of feature interdependencies. The suggested fuzzy clustering technique improves the majority sample dataset within the Advanced Integrated Fraud Detection System (AIFDS) framework by removing outliers. The dataset utilization is thoroughly evaluated using modern classifiers such as XGBoost, CATBoost, Decision Tree and Random Forest. The AIFDS, enhanced by the MVOA-with fuzzy clustering model achieved better performance in accuracy of 94.32%, Sensitivity of 96.37 and precision of 93.49%, and enhancing fraud detection and lowering the financial risks associated with fraudulent operations.