Question

Q2A Construct a decision tree with root node Type from the data in the table below. The first row contains attribute names. Each row after the first represents the values for one data instance. The output attribute is Class. Scale Type Shade Texture Class One One Light Thin A Two One Light Thin A Two Two Light Thin B Two Two Dark Thin B Two One Dark Thin с One One Dark Thin С One Two Light Thin C

Answer 1

Decision tree for the above given table :

TYPE One TWO SCALE SCALE One Two One Two SHADE SHADE SHADE SHADE Dark Light CLASS CLASS CLASS CLASS CLASS CLASS CLASS CLASS An ma un ni ANN AN |||||||| |||||||| A A B A B TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE TEXTURE THIN THIN THIN THIN THIN THIN THIN THIN THIN THIN THIN THIN THIN THIN THIN THIN

Explanation :

Every attribute given in the table has two values.

So, we need two children nodes per attribute nodes.

Given, to take "Type" as root node, we have two value either to take one or two. Taking one decision leads to the second node which is "Scale".

From Scale node, we again take decision based on "One/Two" and go to the next level.

By traversing all the nodes, at last we take "Texture" node as leaf node as it has one single value which is "Thin".

So in this decision tree, no matter what route you take or what decisions you make form selecting the type, you will always end up with thin texture.

For example, One person wants to select a cloth material.

So, at node Type, he selected "one"

Then at Scale node, he selected "two"

At Shade node, he needed light shade so he selected "Light"

At Class node, he found class A as preferable, So he selected "A".

Finally he was left with thin texture so he selected "Thin" as texture.

So, the decision path for this person becomes,

One --> Two --> Light --> A --> Thin