Go backward to 1 Decision-tree Learning
Go up to Top
Go forward to 3 Neural Network

2 Decision Tree Evaluation

Write a program that evaluates binary decision trees. A binary decision tree is either a value or of the form if(Att=Val,T₁,T₂), where T₁ and T₂ are decision trees.

You should assume that all of the data on examples is given using the relation:

prop(Obj,Att,Val)

You need to write a relation:

dteval(Obj,DT,Cl)

that is true if object Obj is classified by decision tree DT as having value Cl.

For example, suppose example e₁ defined by

prop(e_1,a,true).
prop(e_1,b,true).
prop(e_1,c,false).

The query

? dteval(e_1,if(b=true,if(a=true,true,false),if(c=true,false,true)),Val).

has as its answer Val=true.

Axiomatize dteval. You can assume the predicate value(V) that is true if V is a legal value, as well as the predicates prop (as above) and \= (where W \=V means W and V are different values).

Solution to decision tree evaluation