The CPCS networks are multi-level, multi-valued BNs for medicine. They were created by Pradhan et al.  based on the Computer-based Patient Case Simulation system (CPCS-PM) developed by Parker and Miller . Two CPCS networks were used in our experiments. One of them consists of 422 nodes and 867 arcs, and the other contains 364 nodes. They are among the largest BNs in use at the present time.
The CPCS networks contain abundant causal independencies. As a matter of fact, each non-root variable is a convergent variable with base combination operator MAX. They are good test cases for inference algorithms that exploit causal independencies.