Neural network based fuzzy cognitive map

A Fuzzy Cognitive Map (FCM) is a powerful technique for modeling and analyzing complex systems. In this study, we propose a novel learning algorithm that, unlike existing FCM-based learning algorithms, ensures matching the desired system state by computing the otherwise “unexplained” biases in the model. Our learning algorithm considers both the whole system bias and the individual biases for each system factor (concept). We explore the impact of FCM structure and characteristics for the proposed algorithm and suggest an interpretation of computed biases. Finally, we propose an FCM visualization technique which enables comparison between and a deeper understanding of modeled systems. As FCMs offer a broader, quantifiable view of the causal relationships between factors, the approach used in this study provides insights into FCM modeling and application to real-world complex systems.