Certainty Factor Propagation

Rule-based expert systems like MYCIN never assumed evidence was certain — instead every fact and rule carried a certainty factor (CF) on a −1..+1 scale, and the inference engine had exact formulas for combining them: minimum for AND, maximum for OR, and a special non-additive rule for merging two independent lines of evidence pointing at the same hypothesis. This simulator renders a small diagnostic rule network in 3D — three uncertain evidence nodes feeding two rules that both conclude the same hypothesis H — and lets you set each evidence CF, choose whether R1's antecedent is a conjunction or disjunction, and press Propagate to watch a pulse carry each CF value through the network while the live readouts and node colors update with the real MYCIN combination formulas.