Causal and Plausible Reasoning in Expert Systems

This study sets out to establish a unified framework for causal and plausible reasoning. We identify a primitive set of causal roles which a condition may play in the inference. We also extend Dempster-Shafer theory to compose the belief in conclusion by the belief in rules and the belief in conditions. The combined framework permits us to express and propagate a scale of belief certainties in the context of individual roles. Both the causation aspect and the certainty aspect of an inference are now accounted for in a coherent way.