An expert system has four major components:
1. Natural (software) language interface for the user
2. Knowledge base (Like a database, where the facts are stored)
3. Inference machine(software that solves problems and makes logical inferences)
4. Explanation (which explains its conclusions to the user)
One of the most famous expert systems-an older system now being replaced by updated ones is MYCIN, a system that diagnosis infectious disease and recommends appropriate drugs. For example, bacterium can be fatal if it not treated quickly. Unfortunately, tradition tests for it require 24 to 48 hours to verify a diagnosis. However, MYCIN provides physicians with a diagnosis and recommended therapy within minutes.
To use MYCIN, the physician enters data on a patient; as the data is being entered, MYCIN asks questions. As the questions are answered MYCIN's inference machine "reasons" out a diagnosis, which allow relatively unsophisticated users to build their own expert system. Such software tools for building expert systems are called shells. For IBM computers, helps executives plan effective negotiations with other people by examining their personality types and recommending negotiating strategies.
Knowledge base, inference engine, user interface