Expert systems which maintain knowledge about objects whose attributes are time-variant must have an awareness of time. This awareness can be made manifest by incorporating time in the quantification of uncertainty of aging knowledge about such objects.

Many expert systems use some method to quantify the degree of belief, or uncertainty, of their knowledge. Examples of these methods include Bayesian probability theory, certainty factors of EMYCIN, the Dempster-Shafer theory, and fuzzy logic. These methods offer different representations for measures of confidence, and different calculi for combining these measures. We describe an extension to such confidence measures by adding a dimension of time.

We propose the concept of Decaying Confidence Functions to express the time-varying uncertainty of aging knowledge. Decaying confidence functions specify how confidence in knowledge should decrease as the knowledge gets older. We describe how this can lead to efficiencies in expert systems which must deal with time-varying information, such as expert systems used for monitoring real-time systems.




Download Full History