This article aims to introduce the concept of knowledge and its relationship with expert systems. The components of an expert system and the process of acquiring knowledge in its development are also illustrated as one of the most important parts. In many occasions, this acquisition process is not carried out correctly, so here is a recommendation to achieve it.
Introduction
Data, information, knowledge. Each one is the result of the processing and use of the previous one: information is a group of data organized for a purpose within a context and knowledge is a mixture of experience, values and information (Carrión, n / d).
One of the biggest problems in companies today, whether large or small, is the use of knowledge that is constantly generated through all its processes and all its personnel. This difficulty in using the knowledge of the company arises because its sources are very varied and range from documents to databases, discussion groups and the people themselves.
Development
How can knowledge help my company?Each company carries out its daily operations through certain processes: the operation of machinery, the assignment of workload to technicians, customer service by telephone, the creation of a website for the site, the repair of a computer, the purchase of supplies, the process of a guarantee, etc. The simplest case to explain how knowledge can be used in a company is when a new person starts his activities, most of the time he does not have the necessary expertise to carry them out, so he must be trained and given time, sometimes long, to adapt and learn to solve the different problems that arise. If this person had a base that stored the knowledge that arises every day in each person who works in the company, their adaptation process would be easier.But many times, to solve a problem it is not enough to have a knowledge base: you need an expert system.
According to Turban and Aronson (2001), an expert system is "a system that uses human knowledge captured in a computer to solve problems that ordinarily require human expertise." These systems have 5 basic components and are illustrated in the following diagram (WTEC, 1993):
Figure 1. Basic structure of an Expert System
The first step, before developing an expert system, is to understand how the information is received and how it is transformed and used to arrive at the final decision. This feature makes the expert system and its development process completely different from one company to another. The people involved in this previous step are the system development engineer and the experts in the subject to be documented.
Knowledge acquisition
The component that is in contact with the direct source of knowledge is the knowledge acquisition subsystem and for this very reason represents one of the biggest bottlenecks in organizations. Many times it is not only an expert but a group of them who are in charge of developing the system and "emptying" their knowledge at the base. There are many tools and methods to obtain this knowledge, among which are the interview, the observation and the creation of scenarios. The following describes a simple method, presented by Milton (2003):
1. Conduct an interview with the expert (s) to: probe the knowledge to be acquired, know the terminology, the purpose of the knowledge and the system.
2. Create a conceptual diagram derived from the interview results and use it to generate questions that meet the purposes of the system.
3. Conduct another semi-structured interview with the expert using the questions from step 2.
4. Generate the concepts, rules, attributes, values, relationships that emerge from the interviews.
5. Represent these elements in the most appropriate way (Text, diagrams, illustrations, hypertext, anecdotes, etc.)
6. Present the results to the expert and allow him to make changes to the knowledge already captured.
7. Consult with other experts and make appropriate modifications.
An important recommendation is that both the expert on the subject and the person in charge of developing the system in the organization join forces to carry out this initial phase.
With the knowledge acquisition subsystem and the ready knowledge base, other processes must be followed to achieve the elaboration of the rules, the interface with the user and the inference mechanisms, however, the purpose of this article is to highlight the importance of Knowledge acquisition process, since it is the initial step in the development of an expert system and lays the foundation for its correct operation.
With a well-founded Expert System developed since its inception, companies like American Express have managed to avoid losses of $ 27 million dollars in fraud and bad credit, Texas Instruments saves $ 1 million dollars documenting all their purchasing processes and the ROI of IBM of a series of expert systems that cost $ 2.5 million was $ 37.5 million (da Silva, 2000).
Most of the examples documented so far of implementing expert systems are in large companies, however, these types of systems can be easily implemented in small and medium-sized companies, through databases such as Access and tools such as the Internet.
From my point of view, the process of acquiring knowledge and therefore the development of the expert system is continuous and constant. There will always be new things: processes, products, problems, etc. For documenting and for this the organization must be aware of the effort and resources required to achieve it. In my opinion, Expert Systems are a very important part of all the tools that help organizations today make better decisions, based on knowledge.
Bibliography
Carrión, J. (n / d). Difference between data, information and knowledge.
Da Silva, JC (2000). Expert systems hold high potential for designers. Journal of Hydraulics & Pneumatics, 36-38.
Milton, N. (2003) Information on Knowledge Acquisition.
Turban E., Aronson JE (2001). Decision support systems and intelligent systems. Ed. Prentice.
WTEC, (1993). Expert Systems Building Tools: Definitions.
