Rule-based Expert System (for Respiratory Disease Symptom Diagnosis), its Knowledge Base and System Testing Prototype

Posted on by

Rule-based Expert System (for Respiratory Disease Symptom Diagnosis), its Knowledge Base and System Testing Prototype

[[ Techniques used in expert systems ]]

Expert systems are Artificial Intelligence (AI) systems that use knowledge bases to come up with solutions. There are two techniques used in expert systems, and specifically rule-based expert systems, namely;

– forward chaining
– backward chaining

 

Forward chaining is a data-driven approach which adds to the knowledge base any new knowledge it finds. Backward chaining on the other hand is goal driven, with all possible end goals known as dictated by the knowledge base, and whenever new knowledge is found, it is added to the list of goals as opposed to the knowledge base like in forward chaining.

In operation, the two techniques are not separated, but rather used together to come up with an efficient Rule-based Expert System. While backward chaining is used to prove hypothesis, forward chaining is used to see if any other facts can be derived from known and newly stored facts.

For the expert system in our case, a respiratory disease symptom diagnosis system, we need to first state the Rules which will be stored in the Knowledge Base. After knowing the rules, we will test the system by assuming that a user wants to know their diagnosis from their symptoms.

 

[Knowledge Base Rules for Respiratory Disease Symptom Diagnosis]

Goal 1: Influenza
Rule 1: IF “persistent dry cough” AND “feeling of general malaise” THEN “Influenza”

Goal 2: Hayfever
Rule 2: IF “runny nose” AND “sneezing” AND “positive reaction to allergens (dust, pollen)” THEN “Hayfever”

Goal 3: Laryngitis
Rule 3: IF “fever” AND “dry cough” AND “feeling of general malaise” AND “inflamed larynx” THEN “Laryngitis”

Goal 4: Asthma
Rule 4: IF “breathlessness” AND “wheezing” AND “triggered by allergens (dust, pollen)” THEN ” Extrinsic Asthma”
Rule 5: IF “breathlessness” AND “wheezing” AND “triggered by exercise, smoke or respiratory infection” THEN “Intrinsic Asthma”

 

[[ Testing the Rule-based Expert system ]]

To test the system, we will use symptoms that apply for multiple respiratory diseases;

– Influenza & Laryngitis,
– Extrinsic & Intrinsic Asthma

 

[ Diagnosis 1: Influenza OR Laryngitis ]

User: I have a dry cough.
System: Is the cough persistent?

User: Yes, why?
System: You might have Influenza.

System: Do you also have a feeling of general malaise?
User: Yes, so it is Influenza?
System: Not exactly. It might also be Laryngitis.

System: Do you have a fever?
User: Yes.

System: Its, not Influenza, its Laryngitis. You should get a laryngoscopy to confirm that your larynx is inflamed.
User: Thank you.

 

[ Diagnosis 2: Extrinsic Asthma OR Intrinsic Asthma ]

User: I have breathlessness.
System: Do you also experience wheezing?

User: Yes.
System: You definitely have Asthma.

User: Asthma?
System: Yes. Did you get the breathlessness and wheezing after exercising, passing through dust/pollen, smoking or after a respiratory infection?

User: After smoking.
System: Then you have Intrinsic Asthma.

 

Rule-based Expert System (for Respiratory Disease Symptom Diagnosis), its Knowledge Base and System Testing Prototype
Artificial Intelligence | thetqweb