An uncertain rule matching technique in an expert system

We developed a special technique, called equivalent shifting in three dimensions (ES3D) for rule matching in a weather forecasting expert system CME. ES3D is able to describe the evolving procedures of domestic systems fully and exactly in three aspects: strength, position, and concept. ES3D is able to overcome the well-known brittleness of the first generation expert systems, then enhances the capabilities of expert systems adapting to evidences. The idea lying in ES3D can be applied to other similar task domains.<
>