At verification and validation, formal methods complement rather than replace statistical testing and field trials by proving robustness over a domain. The input domain is defined by attributes whose variation must be explicitly bounded, and the criteria of Clause 5 are then applied to domains or partitions of domains. Where only part of the domain can be assessed, the starting point is to delimit the portion where validation is meaningful and informative; any such partition should be justified, and the justification should show why the chosen criterion can assess robustness on that partition. Partitions can be simple bounds on inputs for a regression task, or parameters a validator understands, such as organ volume or background brightness in a medical image.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.