Please enable JavaScript.
Coggle requires JavaScript to display documents.
Knowledge Representations (Production Rules (Advantages (Simple Syntax -…
Knowledge Representations
Object Attribute Value Triplets (OAV)
A common way to represent factual information
Objects may be physical or conceptual
Attributes are general characteristics or properties associated with objects
e.g. Interest rate is an attribute for a bank loan
Specialized case of the semantic-net approach
Production Rules
One of the oldest and most used methods to represent domain knowledge in expert systems
Originally a model of human problem solving (Newell & Simon, 1972)
Experts can express their knowledge in the form of rules
Knowledge represented as if.. then.. production rules
Also called condition-action rules
Components of a rule-based system have the form
if <condition> then <conclusion> or if <condition> then <action>
Advantages
Simple Syntax - Natural
Easy to understand and program
Flexible (easy to add or modify)
Disadvantages
Poor on representing structures descriptive knowledge
Not all knowledge can be expressed as rules
Exhaustive search through all the rules until the first fires - Information about a thing is stored in several (potentially unrelated) rules
Rule based Expert Systems
Inference engine using Forward and Backward Chaining
CLIPS is designed for forward chaining whereas PROLOG performs backgward chaining
The choice of inference engine depends on the type of problem
Diagnostic problems are better solved with backward chaining
Prognosis monitoring and control are better done by forward chaining
Forward chaining
A popular implementation strategy for expert systems
It is a strategy of an expert system to answer the question, "What can happen next?"
The interface engine follows the chain of conditions and derivations and finally deduces the outcome
It considers all the rules and sorts before concluding a solution
This strategy is followed for working on conclusion, result or effect
Forward-chaining Sequential inferencing of rules
Check premises of rules one by one to verify true or false
May take a longer time to discover answer due to ordering of rules
Backward chaining
With this strategy, an expert system finds out the answer to the question. "Why this happened?"
The interface engine tries to find out which conditions could have happened in the past for this result
This strategy is followed for finding out cause or reason
For example, diagnosis of blood cancer in humans
Search is from goal to goal
Start from a possible ultimate goal
Try to prove the premises of the ultimate goal
The premises of the ultimate goal become the subgoals
Search continues with the effort to prove true or false these subgoals by looking at the premises of these subgoals
If all the premises of the subgoals are true, then the subgoals and the goal are true
Limitations
Weaknesses of rule-based systems that inspired the rise of model-based reasoning
Experts may not be able to externalize their experience into clean bits of knowledge that can be encoded into rules
Their knowledge is an accumulation and a combination of years of being exposed to many instances of similar problems (and their subsequent solutions)
