Cambridge A Level Information Technology (ICT): Expert systems (9626) – Practice Questions

These are original questions written for Marlbridge, for revision and practice on this content. They are not reproduced past-paper questions, and they do not replicate the exam’s exact structure, question count or mark tariffs – examination boards hold copyright in their own papers. Use these alongside the official past papers from your board or school.

These questions cover topic 7, Expert systems (section 7.1) of the Cambridge International AS & A Level Information Technology 9626 syllabus for examination in 2025, 2026 and 2027 (version 3). This is AS Level content, part of both the AS Level and the full A Level. It is examined on Paper 1 (Theory), which is based on sections 1–11; Paper 2 (Practical) candidates apply knowledge and understanding from sections 1–7. All questions can be answered on paper. All rule sets are fictional and simplified.

Learn the content first with the study guide and the revision notes. The course hub is Cambridge A Level IT, the 9626 checklist lists every outcome, and you can find your gaps with a free diagnostic.

Questions

1. State the two parts of the knowledge base in an expert system. [2]

2. Describe the role of the inference engine in an expert system. [3]

3. Explain the purpose of the knowledge base editor. [2]

4. Explain why an explanation system is useful in an expert system used for medical diagnosis. [3]

5. An insurance company uses these rules:

R1  IF driver age is under 25 THEN risk level is high
R2  IF risk level is high AND driver has made a claim in the last 3 years THEN premium band is D
R3  IF risk level is high AND driver has made no claim in the last 3 years THEN premium band is C
R4  IF premium band is D THEN offer telematics (black box) policy

A 19-year-old driver who made a claim last year applies.

(a) Using forward chaining, list the rules that fire, in order. [3] (b) State the two conclusions the system gives the customer. [2] (c) Explain why this process is described as data driven. [1]

6. A mining company’s expert system uses backward chaining with these rules:

R1  IF mineral M is likely AND a road is within 20 km THEN recommend test drilling
R2  IF magnetic survey reading is high AND rock samples contain quartz veins THEN mineral M is likely

The goal is “recommend test drilling”. Questions are asked in the order the conditions appear.

(a) State the first question the system asks the user. [1] (b) All answers are “yes”. List the questions asked, in order. [2] (c) At a second site the magnetic survey reading is low. Explain what the system does. [2] (d) Give one reason why backward chaining suits this task. [1]

7. A nature reserve’s identification system uses these rules:

R1  IF animal has six legs THEN animal is an insect
R2  IF animal is an insect AND it has hard wing cases THEN animal is in group G
R3  IF animal is in group G AND wing cases are red THEN species is G2
R4  IF animal is in group G AND wing cases are black THEN species is G3

A volunteer enters: six legs; hard wing cases; black wing cases.

(a) Using forward chaining, state the order in which the rules fire and the species identified. [2] (b) Explain why each question the system asks should test only one feature. [2]

8. Explain, using examples, the difference between data driven and goal driven behaviour of computer-controlled characters in a game. [4]

9. Describe how an expert system could be used to help a family with financial planning. [4]

10. Explain how artificial intelligence used to manipulate social media can be:

(a) data driven [2] (b) goal driven. [2]

11. A garage plans to use an expert system to diagnose car engine faults. Describe how each component of the expert system would be used. [7]

12. A bank lets customers use an expert system for investment analysis instead of meeting an adviser. Evaluate this decision. [6]

Answers

1. A database of facts [1]; a rules base. [1] [2] Examiner insight: “Facts and rules” earns both marks only if both are named; “a database” alone earns nothing because it does not separate the two parts.

2. It takes the facts entered through the user interface and searches the rules base for rules whose conditions match them [1]; it applies (fires) matching rules to reach new conclusions, using forward or backward chaining [1]; it decides which question to ask next and passes the possible solutions to the user interface. [1] [3] Examiner insight: “It is the brain of the system” is too vague for credit; each mark needs a specific action the engine carries out.

3. It allows facts and rules in the knowledge base to be added, edited or deleted [1] so that the system stays up to date when knowledge changes, for example new research or new products. [1] [2] Examiner insight: Answers that say users enter their symptoms through the editor confuse it with the user interface and score zero.

4. It shows how the system reached a diagnosis, listing the symptoms and rules used [1], so the doctor can check the reasoning is sound before acting [1]; it can explain why a question is being asked or show a probability for each possible condition, which helps the doctor and patient trust and judge the advice. [1] [3] Examiner insight: The question is about medical diagnosis, so at least one point must be applied to doctors, patients or diagnoses; a generic definition earns at most one mark.

5. (a) R1 fires first [1], then R2 [1], then R4. [1] (b) Premium band D [1]; offer of a telematics (black box) policy. [1] (c) It starts from the data entered (age and claim history) and works forward to conclusions. [1] [6] Examiner insight: Listing R3 as firing, or omitting R4 because it fires only after R2 adds a new fact, loses the accuracy marks; the order must be exact.

6. (a) Is the magnetic survey reading high? [1] (b) Is the magnetic survey reading high?; do rock samples contain quartz veins? [1]; is a road within 20 km? [1] (in that order) (c) R2 fails at its first condition, so “mineral M is likely” cannot be proved [1]; R1 therefore fails and test drilling is not recommended, without asking about quartz veins or the road. [1] (d) Only the questions needed to test the goal are asked, so costly survey or sampling data is collected only when needed. [1] [6] Examiner insight: In (b) the road question must come last, because backward chaining proves the sub-goal “mineral M is likely” before the second condition of R1.

7. (a) R1, then R2, then R4 [1]; species G3. [1] (b) If a question tests two features, a user whose animal has one but not the other cannot answer it correctly [1], so the inference engine may receive wrong facts and give a wrong identification. [1] [4] Examiner insight: In (a) the species mark depends on the rule order being correct; naming G3 with an order that includes R3 does not earn both marks.

8. A data driven character reacts to the current state of the game [1], for example IF the player comes within range THEN attack. [1] A goal driven character starts with a goal and works backward to plan the actions needed [1], for example to capture an object it must reach a room, so it must first open a door. [1] [4] Examiner insight: Each type needs both a description and an example; two descriptions with no examples is capped at two marks.

9. The family enters income, spending, savings, debts and goals, such as buying a house, through the user interface [1]. The knowledge base holds experts’ rules on budgeting, saving, loans and pensions [1]. The inference engine applies the rules to the family’s data to work out what they can afford to save or borrow [1]. The system outputs a possible plan, such as a monthly savings amount, with an explanation of how it was reached. [1] [4] Examiner insight: “Describe how” needs the input, the processing and the output for this family; a list of component names with no financial content earns little.

10. (a) The system collects the user’s activity data, such as likes, shares and viewing time [1], and works forward from it to infer interests and choose which posts or adverts to show. [1] (b) The system starts from a goal, such as keeping the user online longer or making them see a particular message [1], and works backward to select the content and timing most likely to achieve it. [1] [4] Examiner insight: Each part needs the starting point and the direction of reasoning; describing (a) and (b) with the same example earns credit only once.

11. Indicative points; seven shown:

• User interface: the mechanic answers questions on screen, such as whether the engine cranks, often as yes/no choices [1], and sees the suggested faults. [1]
• Knowledge base: the database of facts holds faults and their symptoms for each model [1]; the rules base holds IF…THEN rules linking symptoms to faults. [1]
• Inference engine: applies the rules to the mechanic’s answers to reach likely faults and decides which question to ask next. [1]
• Explanation system: shows which symptoms and rules led to each suggested fault, so the mechanic can check it. [1]
• Knowledge base editor: lets experts add rules and facts for new car models. [1] [7] Examiner insight: Every component needs a garage-specific use; naming all five components without applying them to car engines earns well below half marks.

12. Indicative points; for each point one mark for the benefit or drawback and its development.

• Available at any time without an appointment, so customers can check investments when they choose. [1]
• Gives consistent analysis based on several experts’ knowledge, not one adviser’s view. [1]
• Cheaper for the bank than employing many advisers. [1]
• Cannot take account of unusual personal circumstances that are not in its rules, so advice may be unsuitable. [1]
• If market rules are not updated through the knowledge base editor, advice goes out of date and could lose customers money. [1]
• Conclusion: suitable for routine analysis, but customers with large or complex investments should still be able to speak to an adviser. [1] [6] Examiner insight: “Evaluate” needs both sides and a reasoned judgement; a one-sided list usually earns no more than about half the marks.

Where marks are usually lost

• Naming “a database” for the knowledge base without its two parts.
• Treating the knowledge base editor as the way users enter data.
• Describing the explanation system as a help screen rather than an account of the reasoning.
• Missing a rule that fires only after a new fact has been added in a forward chaining trace.
• In backward chaining, listing questions in the order they appear on the page rather than the order sub-goals are tested.
• Forgetting that a failed condition stops the rule, so later questions are never asked.
• Giving data driven and goal driven definitions with no example from the context in the question.
• Writing generic advantages (“it is fast”) with no link to the scenario.

Next steps

• Review weak areas in the revision notes
• Re-read the study guide for full explanations
• More AS practice: The digital divide practice and Algorithms and flowcharts practice
• Course hub: Cambridge A Level IT
• Printable checklist: 9626 checklist
• Find your gaps with all free 10-minute diagnostics
• Book a free trial class

Official syllabus

Cambridge International AS & A Level Information Technology 9626 syllabus for examination in 2025, 2026 and 2027, version 3 (published July 2025), Cambridge International, part of Cambridge University Press & Assessment. Topic 7, Expert systems, section 7.1.