Interval-Valued Fuzzy Cognitive Maps with Genetic Learning for Predicting Corporate Financial Distress

Abstract

Fuzzy cognitive maps (FCMs) integrate neural networks and fuzzy logic to model complex nonlinear problems through causal reasoning. Interval-valued FCMs (IVFCMs) have recently been proposed to model additional uncertainty in decision-making tasks with complex causal relationships. In traditional FCMs, optimization algorithms are used to learn the strengths of the relationships from the data. Here, we propose a novel IVFCM with real-coded genetic learning. We demonstrate that the proposed method is effective for predicting corporate financial distress based on causally connected financial concepts. Specifically, we show that this method outperforms FCMs, fuzzy grey cognitive maps and adaptive neuro-fuzzy systems in terms of root mean squared error.