TYPES OF EXPERT SYSTEMS
10.13 TYPES OF EXPERT SYSTEMS
Expert systems appear in many varieties. The following classifications of ES are not exclusive; that is, one ES can appear in several categories.
EXPERT SYSTEMS VERSUS KNOWLEDGE-BASED SYSTEMS
Accordi ng to this classification, an expert system is a system w h o s e behavior is so sophisticated that we woul d call a person w h o p e r f o r m e d in a similar manner an
expert. M Y C I N and X C O N are good examples. Highly trained professionals diagnose
b l o o d diseases ( M Y C I N ) and configure c o m p l e x computi ng equipment ( X C O N ) . These systems truly attempt to emulate the best human experts.
In the commercial world, however, there are systems that can effectively and effi- ciently perform tasks that do not really need an expert. Such systems are called knowl- edge-based systems (also known as advisory systems, knowledge systems, intelligent
job-aid systems, or operational systems). As an example, let us look at a system that gives advice on the immunizations r e c o m m e n d e d for travel abroad. This advice depends on many attributes, such as the age, gender, and health of the traveler and the country of destination. O n e needs to be knowledgeable to give such advice, but o n e n e e d not be an expert. In this case, practically all the relevant k n o w l e d g e is docu- mented in a manual available from most public health departments (in only 1 or 2 per- cent of the cases is it necessary to consult a physician). A n o t h e r example is automated help desks (see A I S in Action 10.8).
The distinction between the two types of ES systems may not be so sharp in reality.
C H A P T E R 1 0 ARTIFICIAL INTELLIGENCE A N D EXPERT SYSTEMS: K N O W L E D G E - B A S E D SYSTEMS 5 6 7
A I S I N A C T I O N 1 0 .8 AUTOMATING THE HELP DESK
Millions of employees work in organizations as vide advice on how to fix problems in the point-of- providers of information and are in direct contact with
sale terminals at the many Color Tile stores. Using the customers. Often customers are frustrated because all ES, operators can now determine the solution to the the lines are busy when they call an information center.
problem much faster and more accurately. Such sys- ("All agents are busy. You are important to us; please
tems are now available for employees on intranets stay on the line. Someone will be with you as soon as
and for customers on extranets. Peppers et al. (1999) possible ") Also, the information provided may not be
provide the example of Canadian Tire Acceptance accurate. The solution is to automate the help desk by
Ltd., which serves 4 million credit card holders. By using expert systems.
employing Web technology the intelligent center inte- An example is Color Tile Company, which uses grated all incoming inquiries (fax, telephone, Web). Expert Advisor (from Software Artistry Inc.) to sup- Using an ES, the system analyzes customers' profiles port queries from its own employees. Formerly, opera- and recognizes needs so that better service can be tors had to search through numerous manuals to pro- provided.
Basically it is a matter of how much expertise is included in systems that classifies them in o n e category or the other. Knowledge systems can be constructed more quickly and
cheaply than expert systems. RULE-BASED EXPERT SYSTEMS
Many commercial expert systems are rule-based systems because the technology of rule-based systems is well developed and the development tools can be used by end- users. In such systems, knowledge is represented as a series of rules.
FRAME-BASED SYSTEMS In frame-based systems, knowledge is represented as frames, a representation of the
object-oriented programming approach (see the discussion of knowledge representa- tion in Chapter 11).
HYBRID SYSTEMS Hybrid systems include several knowledge-representation approaches; at a minimum
they typically involve frames and rules. Advanced techniques such as artificial neural networks and fuzzy logic are sometimes integrated with rules to provide better advices.
MODEL-BASED SYSTEMS Model-based systems are structured around a model that simulates the structure and
function of the system under study. The model is used to compute values that are com- pared to observed values. The comparison triggers action (if needed) or further diag- nosis (Chapter 12).
READY-MADE (OFF-THE-SHELF) SYSTEMS Expert systems can be custom-made to meet the needs of a specific user or purchased as
ready-made packages for general use. Ready-made systems are similar to such application
P A R T I V INTELLIGENT DECISION SUPPORT SYSTEMS
ment. Ready-made systems enjoy the economy of mass production and therefore are con- siderably less expensive than customized systems. They can be used as soon as they are
purchased (several are available on the Web). Unfortunately, ready-made systems are very general in nature, and the advice they render may not be of value to a user involved in a complex situation. However, their popularity has b e e n increasing as their prices
decrease and their capabilities increase. There are two types of ready-made system: those for general use, and those that are industry-, country-, or product-specific.
REAL-TIME EXPERT SYSTEMS In real-time Expert Systems, a strict limit is set on the system's response time, which
must be fast enough to control the process being computerized. In other words, the sys- tem always produces a response by the time it is needed.