On the developpement of an expert system for student's evaluation: case of a network course

I.2. Notion of the expert systems

Professor Edward Feigenbaum of the University of Stanford, pioneer in technology of the expert systems defined it as "a program of intelligent treatment which uses knowledge and the procedures of inference to solve problems which are so difficult that they require a significant human experiment for their solution" (Feigenbaum 82).

Thus, an expert system is a system of treatment which emulates the skill to make decisions of a human specialist.

The verb «emulate» means that the expert system has the objective to act in all the aspects like a human specialist.

The expert systems are a branch of the artificial intelligence which makes full use of knowledge specialized to solve problems like a human specialist. This last is a person who has a knowledge developed in a given field. The specialist thus has special knowledge or skills that the majority does not know, it can solve problems which the majority cannot solve, or it solves them with much more efficiency.

The terms expert systems or system based on knowledge or expert system based on knowledge are used like synonyms.

The I.2 figure illustrates the basic concept of an expert system. It includes two principal components: The base of knowledge contains the knowledge which allows the mechanism inference to make conclusions; these one are answers of the expert system to the special consultation of the user. The following figure is showing basic concepts of functioning of an expert system:

User

Basic knowledge

Inference mechanisme

Facts

Experts

Aptitudes

Figure 1.1 Basic Concepts of functioning of

Expert system

Contrary to the general techniques of the solution of the problems, the knowledge of a specialist specifically concentrates on the field of the problem, like medicine, finances, sciences, engineering, etc, in which the specialist can solve problems with facility. The expert systems are conceived specialized in a field of the problem.

The knowledge of the specialist to solve specific problems is called field of knowledge of the expert.

In its field of knowledge, an expert system reasons or makes inferences in the same way that a human specialist can deduce the solution of a problem. I.e., being given certain facts, we deduce a solution.

I.2.1. Advantages of an expert system

The attractive advantages of expert systems are as follows:

o Great availability: The experiment is available for any equipment of adequate treatment. In a more real direction, an expert system is the massive production of the experiment.

o Reduced cost: The cost to place the experiment at the disposal of the user is enormously reduced.

o Reduced danger: The expert systems can be used in environments which could be dangerous for a human being.

o Permanence: The experiment is permanent. Contrary to the human specialists who can withdraw themselves, give up or die, the knowledge of an expert system will last indefinitely.

o Multiple experiment: The knowledge of several specialists can be available to work simultaneously and continuously on a problem, at any hour of the night or day. The level of experiment combined with many expert systems can exceed that of only one human specialist (Harmon 85).

o Explanation: The expert system can explain clearly and in detail the reasoning which leads to a conclusion that increases the confidence which the decision taken was correct. A human being can be tired, can give up or not to be able to always do it.

o Fast answer: A fast response or in real time can be necessary for certain applications. According to the software or hardware used, an expert system can answer more quickly and to be pus laid out that a human specialist, so that an expert system in real time constitutes a good choice (Hugh 88, Ennis 86).

o Solid, complete answers and without emotions, in any moment: This can be very significant in real time and situations of emergence, when a human specialist does not function with all his capacity because of the pressure and tiredness.

o Intelligent teaching: An expert system can act as an intelligent teacher while leaving that the student carries out programs of example and by explaining the reasoning of the system.

o Intelligent Base data: The expert systems can be used to have access to a data base in an intelligent way.