Diagnosis under compound effects and multiple causes by means of the conditional causal possibility approach

The paper addresses uncertain reasoning on a causal model given by two layered networks, where nodes in one layer express possible causes and those in the other are possible effects. Uncertainty of causalities is expressed in a novel manner, i.e. by Conditional Causal Possibilities. The expression has two advantages over the conventional way with conditional possibilities: it expresses the exact degrees of possibility of causalities, and the number of necessary conditional causal possibilities is far smaller than that of conditional possibilities. However, it has a weakness that it cannot handle causalities with compound effects such as synergistic and canceling effects by multiple causes. The paper discusses the weakness and proposes a solution. First, it discusses how to deal with the compound effects and proposes a new causal model with conditional causal possibilities by multiple causes. Then, it defines a causality consistency problem that calculates possibility of a hypothesis given some observed events, and shows a way to solve the problem.