Expert Systems

An expert system program differs from a conventional program in the way knowledge is presented and processed. This difference exists because:

It attempts to simulate human reasoning. A conventional program has a sequence of predefined instructions for responding to an event. An expert system program on the other hand attempts to reason a response to an event using its knowledge base and logistic rules for problem solving.
It uses representations of human knowledge. These are expressed in a special purpose language such as OPS5 or Lisp. This is in addition to performing standard numerical calculations or data retrieval.
It uses investigative (heuristic) methods to obtain a solution. This is an approximate method that has varying amounts of uncertainty. A conventional program on the other hand uses mathematical and statistical methods to obtain a precise solution.

Thus, comparing an expert system to a conventional program:

	Expert System = Knowledge Base + Reasoning Engine
	Traditional Program = Data + Algorithm

An expert system is often referred to as a knowledge based system. Rather than using numerical methods, a knowledge based system applies 'rules of thumb' to a set of knowledge which is represented symbolically. To illustrate the subtle difference between knowledge based systems and expert systems, one could say that a system able to converse about the weather would be a knowledge based system even if it didn't embody any expertise in meteorology, but an expert system in the domain of meteorology ought to be able to provide weather forecasts. In a knowledge based system, one can say that the representation of knowledge is known as the knowledge base, whereas the part of the program that performs the reasoning is referred to as the inference engine.