Suppose a backward chaining rule based expert system (IDENTIFIER) is constructed

Question

Suppose a backward chaining rule based expert system (IDENTIFIER) is constructed which can recognize various animals in the zoo (see the separate sheet for rules).

a)Simulate IDENTIFIER’s behavior on the given facts. Take care to use the hypotheses in the specified order. Indicate your answer by showing the sequence in which the rules fire and name of the selected animal.

b)Is IDENTIFIER’s ultimate answer sensitive, in any circumstances, to the order of hypotheses? Explain.

IDENTIFIER

Assume the IDENTIFIER pursues hypotheses in the following order:

cheetah

tiger

giraffe

zebra

ostrich

penguin

albatross

Now, here are some facts to assume TRUE, if inquiry is made:

The animal has feathers.

It lays eggs.

It does not fly.

It is black and white.

It swims.

And here are some facts to assume FALSE, if inquiry is made:

It has hair.

It gives milk.

It flies.

It eats meat.

Its eyes point forward.

It has pointed teeth.

It has claws

It has hoofs.

It chews its cud.

It has a brownish color.

It has dark spots.

It has black stripes.

It has long legs.

It has a long neck.

It has a white color.

It is a good flier.

IF-THEN Rules (for Question 1) Rule 10: IF he animal is a carnivore AND it has a brownish color AND it has black stripes Rule 1: LF the animal has hair THEN it is a tiger THEN it is a mammal Rule 2: IF the animal gives milk Rule 11: IF the animal is an ungulate AND it has long legs THEN it is a AND it has a long neck. AND it has a brownish color Rule 3: IF the animal has feathers it has dark spots THEN it is a bird THEN it is a giraffe Rule 4: IF the animal flies AND it lays eggs Rule 12: IF the animal is an ungulate THEN is a bird. AND it has white color Rule 5: IF the animal is a mammal AND it has black stripes AND eats meal THEN it is a zebra THEN it is a carnivore. Rule 6: IF the animal is a mammal Rule 13: IF the animal is a bird AND it does not fly AND it has pointed teeth AND it has long legs AND has claws AND it has a long neck AND its eyes point forward AND it is black and white THEN it is a carnivore. THEN it is an ostrich Rule 7: IF the animal is a mammal Rule 14: IF the animal is a bird AND it has hoofs AND it does not fly THEN it is an ungulate. AND AND it is black and white Rule 8: IF the animal is a THEN it is a penguin AND it chews cud THEN 1) it is an ungulate Rule 15: IF he animal is a bird 2) it is even-toed AND it is a good flyer THEN it is an albatross Rule 9: IF he animal is a carnivore has a brownish color AND it has dark spots THEN t is a

Explanation / Answer

Backward chaining:

Backward chaining works in reverse to forward chaining, and starts from the goal and tries to find data to prove its goal. Therefore, it is also called a goal-driven reasoning. After starting from the given goal, the search of THEN parts of the given rules (action part) (RHS) is conducted, and if the rule is found and its IF part (condition) matches the data in the database, then the rule is executed (fired). Otherwise, if the condition does not match the data (facts) in the database, the inference engine sets the rule that is working on a stack and makes a new subgoal to prove the condition in the current rule. The knowledge base keeps looking for rules to prove the subgoal. The process of stacking the rules is repeated until the knowledge base has no rules to prove the subgoal.

The goal is to decide whether the facts true or false, based on a rule base containing the given rules:

1. If the animal has hair – Then animal is a mammal

2. If the animal gives milk – Then animal is a mammal

3.If the animal has feathers – Then animal is a bird

4. If the animal flies AND lays egg – Then animal is a bird

5. If the animal is a mammal AND it eats – Then animal is a cornivore

and so on....

How does a resolution strategy work in backward chaining? Assume the following rules are available:

R1: IF A AND B AND C THEN D

R2: IF D THEN E

The axioms are given that A, B, and C are true. The goal is to deduce E if a sub-goal D is true.

In order to prove that E is true we need to make sure that the sub-goal D is true. The sub-goal D is proved to be true since the facts A, B, and C are true. So E is deduced from the fact that D is true.

The backward chaining algorithm can be represented in the following form:

1) The rule that matches the goal is selected. IF the condition (IF part) is empty, ask the user for information. ELSE WHILE not end, AND we have the selected rules DO

2) Add the conditions of the rules IF the condition is not met, THEN put the condition as a goal to solve END WHILE.

Related Questions

Navigate

• Browse (All)
• Browse S
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.