Comparative Analysis of Symbolic Reasoning Models for Fuzzy Cognitive Maps

Abstract

Fuzzy Cognitive Maps (FCMs) can be defined as recurrent neural networks that allow modeling complex systems using concepts and causal relations. While this Soft Computing technique has proven to be a valuable knowledge-based tool for building Decision Support Systems, further improvements related to its transparency are still required. In this paper, we focus on designing an FCM-based model where both the causal weights and concepts’ activation values are described by words like low, medium or high. Hybridizing FCMs and the Computing with Words paradigm leads to cognitive models closer to human reasoning, making it more comprehensible for decision makers. The simulations using a well-known case study related to simulation scenarios illustrate the soundness and potential application of the proposed model.