Expert System
An expert system is a computer program that combines a knowledge base of information about a particular field with a system of rules that applies the knowledge to specific situations. Such a system can analyze a problem in a given discipline and provide solutions for it. The system of rules, often called an inference mechanism, consists of a complicated "tree" of logical operations that is carried out by the program. An expert system is considered a form of artificial intelligence. Edward Feigenbaum, a computer specialist, and Joshua Lederberg, a professor of genetics, are generally credited with developing the first expert system at Stanford University in the 1960s. Their program, DENDRAL, determines the structures of organic chemicals using only complex spectroscopic or instrument readings. Other researchers realized that expert systems would be useful in medical diagnosis and began developing programs to assist physicians in investigating disease. An early program of this type, also developed at Stanford University in the mid-1970s, diagnoses blood-related infections.
The program, called MYCIN, analyzes data input by the physician, asks relevant questions about the patient's test results and condition, then uses this information to provide a diagnosis and recommendations for treatment. Although its accuracy seems to be quite good, it has not come into use in medical practice because physicians, fearing serious errors, seem hesitant to rely on a computer to diagnose serious illness. A training program based on MYCIN called GUIDON is, however, being used in medical school classrooms. Another medical expert system, PUFF, has found more use in real-life situations. When a patient undergoes lung tests, PUFF receives the data from the tests as they are carried out and, like MYCIN, provides the physician with a diagnosis and suggested treatment. Expert systems have become very useful in the field of finance and accounting as well. For example, an expert system called TAXMAN has been developed to help the many people bewildered by the complexity of their annual tax returns. Operable on a personal computer (PC), TAXMAN creates a return by prompting the user to input data, performing necessary calculations, and neatly printing the results. A host of other expert systems has been devised, including programs to perform geological assessment of various land sites, determine the cause of diesel locomotive malfunctions, solve difficult mathematical equations, and help sales representatives tailor merchandise orders to a customer's needs, especially in technically complex markets such as that for computer equipment. Finally, there are even expert systems to aid computer programmers in designing and "debugging" other computer systems.
This is the complete article, containing 419 words
(approx. 1 page at 300 words per page).
