An innovative fuzzy neural network based on circular triangular fuzzy numbers for analyzing cardiovascular disease diagnosis problems

The growing complexity of multi-criteria decision-making (MCDM) problems in nonlinear and uncertain environments has created a fuzzy logic need for developing hybrid mathematical approaches capable of effectively capturing vague, imprecise, and cyclic information. Classical fuzzy frameworks and existing aggregation mechanisms frequently have limitations in modeling periodic uncertainty and interdependent decision-making (DM) assessment simultaneously. To handle such environments, the proposed study designs a DM technique based on novel circular triangular fuzzy numbers (Circular-TFNs), Bonferroni mean operators (BMOs), and fuzzy neural network (FN-Network) structures for intelligent analysis under uncertain environments. Initially, the presented model of Circular-TFNs is defined to represent periodic information more fuzzy logicly than the traditional triangular fuzzy numbers (TFNs). Fundamental characteristics and mathematical behaviors of the newly defined approach are thoroughly examined. Moreover, some basic operational laws are established to illustrate an exhaustive algebraic foundation for the proposed Circular-TFNs. Multiple BMOs based on the proposed operational laws are developed using Circular-TFNs, which are crucial to aggregated interrelated DM assessment while considering interactions among decision criteria. Furthermore, a new FN-Network technique is designed using Circular-TFNs and proposed BMOs to further improve intelligent DM capability. The proposed technique consists of input, hidden, and output layers, where expert information is transformed into collective Circular-TFN assessments, aggregated through BMOs, and processed using membership degrees (MSDs) and score functions to achieve final outcomes. The proposed methodology is employed in cardiovascular disease diagnosis problems to show its applicability and effectiveness. The final outcomes show that the newly defined FN-Network successfully handles uncertain and cyclic medical assessment while providing reliable and flexible diagnostic rankings. Moreover, comparative analyses are demonstrated to evaluate the performance and robustness of the proposed technique against several existing fuzzy-based approaches. The comparative results show that the FN-Network based on Circular-TFNs offers improved flexibility, stability, and accuracy in complex medical DM environments.