Hybrid Fuzzy Logic Models for Performance Evaluation in Complex Decision-Making Systems

This dataset is a supplementary Excel dataset prepared to support the article entitled “Hybrid Fuzzy Logic Models for Performance Evaluation in Complex Decision-Making Systems.” The dataset provides structured experimental data for evaluating the performance of fuzzy logic-based hybrid models, particularly the Fuzzy Logic Controller–Genetic Algorithm (FLC–GA) architecture, in complex decision-making environments characterized by uncertainty, nonlinearity, noise, and dynamic data behavior. The dataset includes reconstructed and synthetic experimental records based on the methodological structure and reported results of the study. It covers three application domains: healthcare decision support, financial risk assessment, and industrial control systems. Each domain is represented by structured feature variables, target labels, and evaluation-related information designed to support model comparison and reproducibility analysis. The Excel file contains several sheets, including dataset overview, metadata, metric definitions, model performance results, robustness analysis, cross-validation fold results, data dictionary, and synthetic domain-specific datasets. The performance indicators include accuracy, Root Mean Square Error (RMSE), robustness score under noise-perturbed conditions, computation time, and comparative results among standalone Fuzzy Logic Controller (FLC), conventional machine learning baseline, and the proposed Hybrid FLC–GA model. This dataset is intended for academic research, benchmarking, replication support, methodological validation, and further development of hybrid artificial intelligence models for decision support systems. It may also be useful for researchers working on fuzzy logic, genetic algorithms, interpretable machine learning, intelligent decision-making, uncertainty handling, and sustainable digital transformation. Please note that this supplementary dataset is a reconstructed/synthetic dataset created for documentation, reproducibility support, and Zenodo archiving purposes. It does not represent confidential raw institutional data.