# Certainty Factors

Expert Systems certainty:

• The facts and relationships of the problem area (e.g., the rules) may contain uncertainty. For example, "If these conditions are met, this outcome almost always results. Once in awhile, a different outcome results."
• The user may express doubt in an answer. For example, "I'm almost positive this is the value, however, value-2 is also a possibility."

Many ES's handle uncertainty in rules and user supplied information through numerical certainty factors. Most are based on MYCIN certainty factors. Certainty factors normally range from some negative value to some positive value. For example from -100 to 100. A certainty factor of -100 would represent a complete lack of belief in something while a factor of 100 would represent an absolute belief in a rule or value.

## MYCIN-like certainty factors

In the equations from this link, CF (previous) is defined as the certainty factor associated with the parameter's value before the action of the THEN portion of a rule is performed. CF is the certainty factor of the new evidence added by a rule.

## CLIPS Rules for Combining MYCIN-like Certainty Factors

A file with these rules can be copied for use from ~engelb/565/cf.clp