Sorry, but copying text is forbidden on this website!
According to Ford (1991), an Expert system is an artificial intelligence technique that uses expert knowledge to solve complex decision problems which can be viewed as a computer simulation of a human knowledge. According to Engelmore and Feigenbaum (1993), the process of building such a system is known as knowledge engineering and its practitioners following this system are called knowledge engineers (Engelmore & Feigenbaum, 1993). The function of the systems is explained as below:
‘’Such a system may completely fulfil a function that normally requires human expertise, or it may play the role of an assistance to a human decision makers’’ (Jackson, 1990, p3). The expert system differs from human expertise in which it allows integration of knowledge from different sources and it make it available for a larger number of people and who are less skilled (Finlay and Dix, 1996 and Ford, 1991). Moreover, the expert systems are usually used commercially because it is less expensive in terms of searching for knowledge and in training new unskilled people.
Finally, expert systems equip the users with unbiased response and thus provide correct solution to problems. “Expert systems not only possess human knowledge in the form of coded tables, databases, and programmed logic, but are coming closer and closer to adequately and truly representing human systems that think. Built to include the ever popular modular structure, expert systems can be refined and improved, just as a human’s thoughts can”, (Ruchelsman, 2004, p. 2)
According to Engelmore & Feigenbaum (1993), the experts systems provide major costs saving for a company which is a result of quality improvement (Engelmore & Feigenbaum, 1993). For example, savings are sometimes in the tens or hundreds of thousands of dollars in small systems; but for large systems, often in the tens of millions of dollars and taken 2. 8. 1 Developing an expert system 2. 8. 1. A. Test phase In order to be certain that the problem will be solved appropriately by an expert system, two tests should be passed, as seen in table 1 (Finlay and Dix, 1996).
First, the problem should fall in one of categories that are shown in table 2. Second, point which should be considered is whether this problem can be sufficiently solved using conventional and cheaper techniques? For example, an analysis should be made to check if the problem can be solved statistically. If the answer to both of these questions is no, then the analysts should consider whether the problem justifies the expenses and effort required to build the required system. This is because one of the main aims of building an expert system is to save costs in the long term.
Building phase When both tests given above are successfully passed by the given problem, several phases are then followed to build such a system. Finlay and Dix (1996), state that the first phase is knowledge base acquisition and they consider it to be the most crucial stage in building the system. Engelmore & Feigenbaum (1993) have mentioned earlier that the knowledge base acquisition contains both factual and heuristic knowledge. According to then the factual knowledge is widely shared.
It is typically called secondary source knowledge and is found in books or journals. While thy say that the Heuristic knowledge is the main expert source knowledge which is rarely discussed, and is more experiential and more judgmental knowledge of performance than the factual knowledge. They call this second phase as Knowledge representation which formalises and organises the previous collected knowledge (Engelmore & Feigenbaum, 1993). Production rule or simply rule is one commonly used representation. A rule consists of a condition statement of IF and THEN parts.
If the IF part of the rule is fulfilled, THEN part can be concluded, or solving the problem action is taken This process is then followed by the design and development phase. According to Lukasheh et al. (2001), this phase consists of a highly iterative set of processes in which the designer builds a part of the system and then tests the result. Here, the system knowledge is modified based on test results and modifications takes place after each single test. In the final phase called the evaluation process, the system is implemented.