EXPERT SYSTEM
Dr. Khoerul Anwar, S.T.,M.T
STMIK Pradnya Paramita
WHAT IS EXPERT SYSTEM
Expert system
is one of the
sub-discipline
(Turban et al 2001) of artificial intelligence
An expert system also known as
knowledge
based system
is a
computer program
that
contains the
knowledge and analytical skills
of one or more
human experts
in a specific
problem domain.
The goal
of the design of the expert system is
to
capture the knowledge of a human expert
relative to some
specific domain
and code
this in a
computer
in such a way that the
CHARACTERISTIC
Expert
Non Expert
Memiliki pengetahuan
{
mengerti,
mempraktekan,
menjelaskan
} banyak
hal (detail) tentang
hal khusus
Sedikit tau tentang
banyak hal
SPECIFIC DOMAIN
2
agriculture
Rice pests
3
fishery
4
computer
Network
DEFINITIONS
Goodall’s (1985) : “An expert system is a
computer system that uses a
representation of human expertise in a
specialist domain in order to perform
unctions similar to those normally
WHEN TO USE AN EXPERT
SYSTEM
If great costs are involved when making the decision
When applied to repetitive problem domains, the task
must be performed often
When a big difference exists between the best solution
and the worst solution
When test data is easy available to test and validate the
ES
When there is a general agreement on the system’s
conclusions. Errors in the input should be tolerated in the
system output. Experts must agree on the solutions.
When recognised expertise is available and a necessity
during the development of the ES
If the problem is clearly specifiable in the area of
CONT..
If the problem is identifiable with a human expert, that is
based on human expertise. Experts must perform the task
substantially better than non-experts. The human
expertise must be scarce.
If the problem domain is well bounded: for example
defining criteria to determine the subject matter within the
system from the matter outside of the system: The task is
reasonably stable and within an acceptable narrow domain.
If solving the problem is based on knowledge rather than
common sense reasoning: ES is based on knowledge and
naïve systems on reasoning.
The solution has to be possible, justifiable and concisely
generated (using only a few hundred rules), and
The knowledge has to be of a cognitive style and
ES IN SOLVING PROBLEMS
Interpretation
Prediction
Diagnosis
Debugging
Design
Planning
Monitoring
Instruction, and
ADVANTAGES OF USING
AN EXPERT SYSTEM
It improves the quality of the system
It can handle uncertainties expressed as
probabilities
It explains the logic behind its
recommendations, making the knowledge
explicit
It can be used as a training vehicle for
users who lack the expertise
It provides monetary savings once
implemented.
Experts are freed enabling them to focus on
CONT…
It can codify and preserve knowledge of
the specific problem domain.
It increases programmer productivity.
It can discover new knowledge
It can provide increased output and
productivity.
It eliminates the need for expensive
equipment used by human experts for
monitoring and control
It makes knowledge and information
CONT…
It can outperform human experts
because of the fact that ES:
Make fewer errors
Do not become tired or bored and never
sleeps.
Will not overlook a solution
Can handle large volumes of data
Can respond more rapidly
Can function in hostile environments such as
deep-sea drillings and reactor control, and
When the knowledge of several experts is
COMPONENTS
Knowledge base :
contains the knowledge necessary for understanding, formulating and for
solving problems.
A declarative representation of the expertise; often in IF THEN rules,
frames, logic, semantic net and case etc.
Both factual and heuristic knowledge (less rigorous, more experiental,
more judgmental knowledge of performance, rarely discused and large individual)
Working storage :
The data which is specific to a problem being solved
Inference engine :
the code at the core of the system which derives recommendations from
the knowledge base and problem specific data in working storage. Inference Engine is a brain of expert system (Here two approaches are used i.e. forward chaining and backward chaining).
User Interface :
the code that controls the between the user an the system
provides facilities such as menus, graphical interface etc. to make the