Abstract
Fuzzy Cognitive Maps (FCMs) are a class of discrete time Artificial Neural Networks that use the structure of recurrent neural networks to create models and make inferences. Contrary to classical Expert Systems, FCM allow cycles and this increases its inference capability. Positive and Negative cycles are defined that reinforce or counteract respectively any initial change introduced in the system. The elementary cycle is that between two neurons. The dynamical behaviour of such a two neuron Fuzzy Cognitive Map using several types of transfer functions is studied. An analysis of their dynamical behaviour is given and the criteria that determine its behaviour are established. A comparison between their dynamical behaviour is also presented.
∗Supported by the Greek National Scholarship Foundation.
∗Supported by the Greek National Scholarship Foundation.
Notes
∗Supported by the Greek National Scholarship Foundation.