6 ICT Applications – Expert Systems
Objective (AO1)
Know and understand how an expert system is used to produce possible solutions for different scenarios.
1. What is an Expert System?
An expert system is a software application that uses artificial intelligence to emulate the decision‑making ability of a human expert in a specific domain. It stores specialised knowledge and applies logical reasoning to generate recommendations, diagnoses or solutions.
2. Core Components (mandatory for the syllabus)
| Component | Purpose | Typical ICT Example |
|---|
| Knowledge Base | Contains facts, rules, heuristics (practical shortcuts) and relationships of the problem domain. | Rules that link network error codes to possible faults. |
| Inference Engine | Applies reasoning (forward or backward chaining) to the knowledge base to draw conclusions. | Engine that matches a printer’s error message with a repair rule. |
| User Interface (UI) | Allows the user to enter data and to receive the system’s conclusions and explanations. | Web form where a technician enters observed symptoms. |
2.1 Optional / Advanced Components (useful but not required for the exam)
- Explanation Facility – shows the reasoning path (e.g., “Because symptom X and Y match fault Z”).
- Knowledge Acquisition Module – tools that help experts add, delete or modify rules, often via interview or questionnaire.
3. How an Expert System Works (AO2)
The system follows a simple cycle, broken down into three numbered stages to match the syllabus format.
- 3.1 User Input – The user enters data about the problem (symptoms, error codes, financial figures, etc.).
- 3.2 Reasoning – The inference engine processes the input using a reasoning method (forward or backward chaining).
- 3.3 Output – One or more possible solutions are presented, often with an explanation of the reasoning.
3.2.1 Reasoning Methods
| Method | Process | Typical ICT Use |
|---|
| Forward Chaining | Start with known facts → apply rules → generate new facts → repeat until a solution is reached. | Network‑monitoring system that begins with logged error messages and derives the likely fault. |
| Backward Chaining | Start with a hypothesised goal (possible fault) → work backwards to see if the required facts are present. | Troubleshooting wizard that first assumes “printer jam” and then checks for jam‑related symptoms. |
3.2.2 Comparison of Forward vs. Backward Chaining
| Aspect | Forward Chaining | Backward Chaining |
|---|
| Typical Speed | Fast when many facts are known. | Can be slower if many possible goals must be tested. |
| Goal Orientation | Data‑driven; discovers all conclusions that follow from the data. | Goal‑driven; works only towards the specific hypothesis. |
| Best Use | Diagnostic systems with many observable symptoms. | Decision‑support where a particular outcome is being investigated. |
3.2.3 Simple Pseudo‑code (forward chaining)
IF symptom = "no‑power" AND device = "router"
THEN possibleFault = "power‑supply failure"
IF symptom = "high latency" AND device = "router"
THEN possibleFault = "congestion"
...
OUTPUT possibleFault + confidence level
4. Example Scenarios (ICT‑focused)
- Network Fault Diagnosis – Input error logs and device status → system suggests router reset, firmware upgrade or cable replacement.
- E‑learning Tutor – Input student quiz results → system recommends personalised study resources and practice questions.
- Business Decision Support – Input sales figures, stock levels, market trends → system proposes ordering quantities or promotional actions.
- Computer Hardware Troubleshooting – Input BIOS beep codes → system outlines step‑by‑step hardware checks.
- Medical Diagnosis (non‑ICT but exam‑relevant) – Input patient symptoms and test results → system suggests possible conditions and further investigations.
5. Benefits of Using Expert Systems (AO3)
- Provides expert advice 24 hours a day, reducing dependence on scarce specialists.
- Ensures consistency – the same input always yields the same reasoning and output.
- Preserves knowledge that might otherwise be lost when experts retire.
- Speeds up decision‑making and can reduce human error.
- Improves user confidence when the explanation facility shows how a conclusion was reached.
6. Limitations (AO3)
- Knowledge base must be regularly updated; outdated rules give incorrect advice.
- Complex, ambiguous or novel problems may lie outside the system’s capability.
- High initial development cost – especially the acquisition and encoding of expert knowledge.
- Risk of over‑reliance – users may ignore professional judgement or contextual factors.
- Version control is essential; unauthorised changes to the knowledge base can cause harmful advice.
7. Evaluating an Expert System (AO3 Checklist)
When deciding whether an expert system is suitable for a particular problem, consider the following criteria. Each links directly to AO3 (analyse, evaluate, make reasoned judgements).
| Criterion (AO3) | Key Questions to Ask |
|---|
| Accuracy & Reliability | Does the knowledge base cover the required domain comprehensively? Are success rates documented? |
| Maintainability | How easy is it to add, delete or modify rules when technology or regulations change? Is version control in place? |
| Cost‑Benefit | Do the time‑saving benefits outweigh the development and ongoing maintenance costs? |
| Ethical & Legal Issues | Is the system handling personal or sensitive data? Does it comply with data‑protection legislation? |
| User Acceptance | Will users trust the system? Is the UI intuitive and are explanations clear enough to build confidence? |
8. Link to the Systems Development Life‑Cycle (SDLC)
Expert‑system projects follow the same SDLC stages as other ICT applications:
- Analysis – Identify the problem, gather expert knowledge, define required rules.
- Design – Choose a reasoning method, design the UI and decide on optional components.
- Implementation – Encode rules into the knowledge base, develop the inference engine.
- Testing – Verify that the system reaches correct conclusions for a range of test cases.
- Evaluation – Use the AO3 checklist to judge suitability, reliability and cost‑effectiveness.
- Maintenance – Update rules, add new knowledge, monitor performance and manage version control.
9. Safety, e‑Safety & Data Protection
- Data Privacy – Systems that process personal health, financial or student data must store and transmit it securely (encryption, access control).
- Integrity of Knowledge Base – Unauthorised changes to rules could cause harmful advice; use version control and audit trails.
- Reliance Risks – Users should be reminded that the system provides advice, not a final decision; professional oversight remains necessary.
10. Summary Checklist (Revision)
- State the definition of an expert system and name the three mandatory components.
- Explain forward chaining and backward chaining, giving one ICT‑focused example for each.
- List at least two real‑world ICT scenarios where expert systems are used (include a non‑ICT example such as medical diagnosis).
- Give one benefit and one limitation; mention a factor to evaluate (e.g., accuracy, cost‑benefit, user acceptance, data protection).
- Recall the link to the SDLC and the purpose of the explanation facility.
11. Sample Examination Questions
- Define an expert system and list its three mandatory components.
- Describe forward chaining and backward chaining. Provide an ICT example for each method.
- Explain how an expert system could be used to diagnose a network router that repeatedly loses connectivity.
- Discuss two advantages and two disadvantages of using an expert system in a business decision‑support context, including at least one ethical or data‑protection consideration.
Suggested Diagram (for revision)
Flowchart: User Input → Inference Engine (forward/backward chaining) → Output with Explanation → (optional) Evaluation & Update.