A document presents equations for scoring rules in a diagnostic and/or prognostic artificial-intelligence software system of the rule-based inference- engine type. The equations define a set of metrics that characterize the evaluation of a rule when data required for the antecedence clause(s) of the rule are missing. The metrics include a primary measure denoted the rule completeness metric (RCM) plus a number of subsidiary measures that contribute to the RCM. The RCM is derived from an analysis of a rule with respect to its truth and a measure of the completeness of its input data. The derivation is such that the truth value of an antecedent is independent of the measure of its completeness. The RCM can be used to compare the degree of completeness of two or more rules with respect to a given set of data. Hence, the RCM can be used as a guide to choosing among rules during the rule-selection phase of operation of the artificial-intelligence system.

This work was done by Mark James of Caltech for NASA's Jet Propulsion Laboratory.

The software used in this innovation is available for commercial licensing. Please contact Karina Edmonds of the California Institute of Technology at (818) 393-2827. Refer to NPO-42717.


This Brief includes a Technical Support Package (TSP).
Equations for Scoring Rules When Data Are Missing

(reference NPO-42717) is currently available for download from the TSP library.

Don't have an account? Sign up here.