Scribd
Upload
36 views12 pages

Lecture Notes 4 Expert Systems

Knowledge acquisition is essential for developing expert systems, involving the gathering and encoding of domain-specific knowledge. Key sources include human experts, databases, and simulations, while methods such as interviews and machine learning aid the process. Challenges like tacit knowledge and evolving domains must be addressed to create effective systems in various fields.

Uploaded by

johnkoolleshan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
36 views12 pages

Lecture Notes 4 Expert Systems

Knowledge acquisition is essential for developing expert systems, involving the gathering and encoding of domain-specific knowledge. Key sources include human experts, databases, and simulations, while methods such as interviews and machine learning aid the process. Challenges like tacit knowledge and evolving domains must be addressed to create effective systems in various fields.

Uploaded by

johnkoolleshan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

MASAAI MARA UNIVERSITY COLLEGE

DEPARTMENT OF COMPUTING & INFORMATION SYSTEMS

COM 4252-1: EXPERT SYSTEMS


Week 4: Knowledge Acquisition in Expert Systems

B Y:
P R O F. S I LVA N C E O . A B E K A
[email protected]
Lecture Outline:
1) Introduction to Knowledge Acquisition
2) Sources of Knowledge for Expert Systems
3) Methods of Knowledge Acquisition
4) Knowledge Engineering Process
5) Case Studies and Real-World Applications
6) Challenges in Knowledge Acquisition
1. Introduction to Knowledge Acquisition
Definition:
Knowledge acquisition refers to the process of gathering, organizing, and encoding domain-specific
knowledge into an expert system.
Importance:
Ensures the expert system's accuracy and effectiveness.
Captures valuable insights from experts and other data sources.
2. Sources of Knowledge for Expert Systems
◦ Human Experts:
◦ The primary source of domain knowledge.
◦ Example: Doctors for medical expert systems, lawyers for legal systems.
◦ Databases and Documents:
◦ Extract knowledge from structured and unstructured data.
◦ Example: Medical records, scientific research papers.
◦ Observations and Simulations:
◦ Analyze real-world phenomena or simulated scenarios.
◦ Example: Observing patient symptoms or simulating chemical reactions.
3. Methods of Knowledge Acquisition
1. Interviews:
◦ Definition: Directly engaging with human experts to extract knowledge.
◦ Types:
◦ Structured (specific questions).
◦ Unstructured (open-ended discussions).
◦ Advantages:
◦ Allows detailed exploration of expert reasoning.
◦ Case Study:
◦ Developing a diagnostic tool for automobile repair using interviews with mechanics.

2. Questionnaires:
◦ Definition: Using surveys to collect knowledge from multiple experts.
◦ Advantages:
◦ Scalable for large groups of experts.
◦ Real-World Application:
◦ Risk assessment in financial institutions using input from various risk managers.
3. Protocol Analysis:

Definition: Observing experts as they perform tasks and verbalize their reasoning.
Example: Monitoring a doctor diagnosing a patient in real-time.
Advantages:
Captures practical and procedural knowledge.
Case Study:
Developing a surgical assistant expert system by observing surgeons.

4. Machine Learning:
Definition: Using algorithms to derive knowledge from data.
Techniques:
Supervised Learning: Labeled data.
Unsupervised Learning: Patterns in unlabeled data.

Advantages:
Automates the acquisition of large-scale knowledge.
Real-World Application:
Fraud detection systems learning from transaction data.
4. Knowledge Engineering Process
Steps:
1. Identify domain and scope.
2. Collect knowledge using methods like interviews or data mining.
3. Organize knowledge into a structured form (e.g., rules or frames).
4. Validate and verify knowledge accuracy.
5. Implement the knowledge into the expert system.
Key Roles:
Knowledge Engineer: Mediates between the expert and the system.
◦ Domain Expert: Provides the raw knowledge.
5. Case Studies and Real-World Applications
Case Study 1: Expert System for Diagnosis in Agriculture
◦ Domain: Crop disease diagnosis.
◦ Description:
◦ Knowledge acquired through interviews with agricultural scientists and farmers.
◦ Built an expert system to recommend treatments for common crop diseases.
◦ Impact: Improved yield and reduced pesticide misuse.

Case Study 2: Fraud Detection System in Banking


◦ Domain: Financial fraud detection.
◦ Description:
◦ Used machine learning to acquire knowledge from transaction data.
◦ Built a system to detect unusual patterns indicating fraud.
◦ Impact: Reduced financial losses due to fraud.
Real-World Application 1: MYCIN
◦ Description: Acquired medical knowledge from experts for bacterial infection diagnosis.
◦ Impact: Pioneered knowledge acquisition techniques for medical expert systems.

Real-World Application 2: IBM Watson


◦ Description: Combines expert interviews and machine learning to process vast medical literature for
cancer treatment recommendations.
◦ Impact: Enhanced precision in personalized medicine.
6. Challenges in Knowledge Acquisition
1. Tacit Knowledge:
• Experts may struggle to articulate their reasoning.

2. Expert Availability:
• Experts may have limited time to participate.

3. Dynamic Knowledge:
• Domain knowledge may evolve, requiring frequent updates.

4. Knowledge Validation:
• Ensuring the correctness and completeness of acquired knowledge.
Summary of the Lecture
Knowledge acquisition is a critical phase in developing expert systems, as it determines the
system's quality and relevance. Various sources, including human experts, databases, and
simulations, provide input, while methods like interviews, protocol analysis, and machine learning
facilitate the process. Despite challenges such as tacit knowledge and dynamic domains, effective
knowledge engineering can yield impactful systems in fields such as agriculture, healthcare, and
finance.
Revision Questions
1) Define knowledge acquisition and explain its importance in expert systems.
2) Describe three primary sources of knowledge for expert systems.
3) Explain the protocol analysis method of knowledge acquisition and provide an example of its
application.
4) Discuss the role of machine learning in knowledge acquisition with a real-world application.
5) What are some challenges faced during knowledge acquisition, and how can they be
addressed?

You might also like