Matakuliah
Tahun
Versi
: H0383/Sistem Berbasis Pengetahuan
: 2005
: 1/0
Pertemuan 13
Membangun Expert System
1
Learning Outcomes
Pada akhir pertemuan ini, diharapkan mahasiswa
akan mampu :
• Merangkaikan tahapan-tahapan dalam
membangun expert systems
2
Outline Material
• Fasa/Tahapan membangun expert
systems
• Beberapa syarat/pertimbangan untuk
membangun expert systems yang baik
3
Phase I: Project Initialization
•
•
•
•
Problem definition
Needs assessment
Evaluation of alternative solutions
(availability of experts, education and training,
packaged knowledge, conventional software)
• Verification of an expert systems approach
(requirements, justification, appropriateness):
• Consideration of managerial issues (project
initiator, financing, resources,legal and other
constrains, selling the project, identifying a
champion)
• Approval of the project in principle
4
12 Requirements to ES
1. Task does not require common sense.
2. Task requires only cognitive, not physical
skills.
3. At least one genuine expert exists.
4. Experts can articulate their methods to
solve problems.
5. Experts agree on the knowledge and the
solution approach to the problem.
6. Task is not too so difficult.
5
12 Requirements to ES
7. Well understood and defined clearly.
8. Task definition fairly stable.
9. Conventional (algorithmic) computer
solution technique are not satisfactory.
10. Incorrect or non optimal result can be
tolerated.
11. Data and test cases are available.
12. The task’s vocabulary has no more than
a couple of hundred concept.
6
8 Justification to ES
Development
1.
2.
3.
4.
5.
6.
7.
8.
Solution to the problem has a high payoff.
ES preserve human expertise not to change.
Expertise is needed in many location.
In hostile and hazardous environments.
Expertise improve performance and quality.
System can be used for training.
ES solution can be derived faster than human.
ES is more consistent and accurate than a
human.
7
Appropriateness of the ES:
• Nature of the problem: problem have
symbolic structure and solution is
heuristic. The task is decomposable.
• Complexity: neither too easy nor too
difficult for human expert
• Scope of problem : manageable size and
should have some practical value.
8
Phase II: System Analysis and
Design
•
•
•
•
•
Conceptual design and plan
Development strategy
Source of knowledge
Selection Computing resources
Feasibility study:
– Economic: cost of system development, maintenance,
cash flow and risk analysis
– Technical: interface requirement, networking issues,
availability of knowledge and data, knowledge
representation scheme.
– Operational: availability of human and other source.
Needs assessment.management and user support.
9
Phase II: System Analysis and
Design
• Cost benefit analysis
• Ending milestone: approved complete
project plan
10
Phase III: Rapid Prototyping
•
•
•
•
Building a small prototype
Testing, Improving, Expanding
Demonstrating and analyzing feasibility
Completing Design
11
Phase IV: System Development
• Building Knowledge Base
• Testing, Evaluating, and improving
Knowledge base
• Planning for Integration
12
Phase V: Implementation
•
•
•
•
•
•
•
Acceptance by Users
Installation, Demonstration, Deployment
Orientation, Training
Security
Documentation
Integration
Field Testing
13
Phase VI: Post Implementation
• Operation
• Maintenance and upgrades
• Periodic evaluation
14
Penutup
• The ES should be developed to fulfill a
recognized and important need.
• The processing speed of the system
should be very high.
• The ES should be able to increase the
expertise of the user
• Error correction should be easy to perform
• The program should be able to respond to
simple question
15
Penutup
• The system should be capable of asking
question to gain additional information.
• Program knowledge should be easy
modified.(add, delete , modify rules)
• The user should fell that they are in control
• Input requirement should be clear and
simple to obtain.
16