Cognitive Maps and Bayesian Networks for Knowledge Representation and Reasoning

Cognitive maps are powerful graphical models for knowledge representation. They offer an easy means to express individual's judgments, thinking or beliefs about a given problem. However, drawing inferences in cognitive maps, especially when the problem is complex, may not be an easy task. The main reason of this limitation in cognitive maps is that they do not model uncertainty with the variables. Our contribution in this paper is twofold : we firstly enrich the cognitive map formalism regarding the influence relation and then we propose to built a Bayesian causal map (BCM) from the constructed cognitive map in order to lead reasoning on the problem. A simple application on a real problem is given, it concerns fishing activities.