Information Technology Reference
In-Depth Information
allows shades of gray, or what is known as “fuzzy sets.” Fuzzy logic rules help computers
evaluate the imperfect or imprecise conditions they encounter and make educated guesses
based on the probability of correctness of the decision.
rule
A conditional statement that links
conditions to actions or outcomes.
•
Using rules.
A
rule
is a conditional statement that links conditions to actions or outcomes.
In many instances, these rules are stored as
IF-THEN statements
, such as “If a certain
set of network conditions exists, then a certain network problem diagnosis is appropriate.”
In an expert system for a weather forecasting operation, for example, the rules could state
that, if certain temperature patterns exist with a given barometric pressure and certain
previous weather patterns over the last 24 hours, then a specific forecast will be made,
including temperatures, cloud coverage, and wind-chill factor. Figure 11.10 shows how
to use expert system rules in determining whether a person should receive a mortgage
loan from a bank. These rules can be placed in almost any standard program language
discussed in Chapter 4 using “IF-THEN” statements or into special expert systems shells
and products, discussed later in the chapter. In general, as the number of rules that an
expert system knows increases, the precision of the expert system also increases.
IF-THEN statements
Rules that suggest certain
conclusions.
Figure 11.10
Mortgage Application for Loans from $100,000 to
$200,000
Rules for a Credit Application
If there are no previous credit problems and
If monthly net income is greater than
4 times monthly loan payment and
If down payment is 15% of the total value of the
property and
If net assets of borrower are greater than
$25,000 and
If employment is greater than three years at
the same company
Then accept loan application
Else check other credit rules
•
Using cases.
An expert system can use cases in developing a solution to a current problem
or situation. This process involves (1) finding cases stored in the knowledge base that are
similar to the problem or situation at hand and (2) modifying the solutions to the cases
to fit or accommodate the current problem or situation. For example, a company might
use an expert system to determine the best location for a new service facility in the state
of New Mexico. The expert system might identify two previous cases involving the
location of a service facility where labor and transportation costs were also important—
one in the state of Colorado and the other in the state of Nevada. The expert system can
modify the solution to these two cases to determine the best location for a new facility
in New Mexico.
