This assignment calls on you to propose a problem area that might be amenable to expert system solution. To do this, you will need to describe the problem area and then show that it plausibly translates into the form that an expert system can solve. You can think of this as a mental stretching exercise to prepare you for thinking about project possibilities.

More specifically, write a brief (approximately two pages, single spaced) paper outlining a problem that you believe is well-suited to solution by expert system. Your paper should include the following elements:

Background

Describe the problem domain well enough that a person unfamiliar with it would be able to follow the rest of your paper and decide if it makes sense.

Expertise

One of the important aspects of an expert system is that it captures expertise: specialize knowledge of a human expert that attacks a particular problem domain. This is distinct from more general, “common sense” knowledge of the world in general. Justify that problems in your domain are solvable without use of such higher-level reasoning (or, alternatively, identify the set of specialized knowledge that attacks problems in your domain). Where could a knowledge engineer access such knowledge (books, scholarly papers, human experts, etc)?

Declarative and Inferential Knowledge

To translate expertise into an expert system, we need to identify declarative knowledge — facts — and inferential knowledge — rules. Give examples of each. If necessary, you may use simplified examples to illustrate your point that the kind of expert knowledge lends itself to such a formalism.

Algorithmic Unsuitability

Just because a problem can be solved by an expert system doesn’t mean it should be. Explain how a pure algorithmic approach would be unsuitable, perhaps because it would be unwieldy, because there is no good algorithmic solution, or because the changing nature of the problem area would make such an approach unmaintainable.