Decision-tree Visualization

Sash Saseetharran
Pre-Requisites

• Basic understanding of data exploration,

machine learning, and Python
Learning Objectives

• Visualizing a decision tree by hand

• Running a test case
Agenda

• Part I
 Explore data
 Visualize
• Part II
 Let’s do it together
• Part III
 Students try in pairs
• Part IV
 Galvanize the understanding
 Closing
Part I
Visualize the tree
Predict if Peter will Play Golf?
Day Outlook Temp Humidity Windy Play
1 sunny hot high false No
2 sunny hot high true No
3 overcast hot high false Yes
4 rain mild high false Yes
5 rain cool normal false Yes
6 rain cool normal true No
7 overcast cool normal true Yes
8 sunny mild high false No
9 sunny cool normal false Yes
10 rain mild normal false Yes
11 sunny mild normal true Yes
12 overcast mild high true Yes
13 rain mild high true No
14 rain hot normal false
Predict if Peter will Play ... Contd

From the given Golf data:

 13 training samples
 One test case
 Within the training data, 8 Yes & 5 No
 Independent Predictor variables:
• Outlook, Temperature, Humidity & Windy
 How many levels?
• 2 levels: Humidity, Windy & Play
• 3 levels: Outlook & Temp
Predict if Peter will Play Golf?
Day Outlook Temp Humidity Windy Play
1 sunny hot high false No
2 sunny hot high true No
3 overcast hot high false Yes
4 rain mild high false Yes
5 rain cool normal false Yes
6 rain cool normal true No
7 overcast cool normal true Yes
8 sunny mild high false No
9 sunny cool normal false Yes
10 rain mild normal false Yes
11 sunny mild normal true Yes
12 overcast mild high true Yes
13 rain mild high true No
14 rain hot normal false
Predict if Peter will Play Golf?
Terminal Node
(Leaf)

Root Node
Predict if Peter will Play Golf?
Day Outlook Temp Humidity Windy Play
1 sunny hot high false No
2 sunny hot high true No
3 overcast hot high false Yes
4 rain mild high false Yes
5 rain cool normal false Yes
6 rain cool normal true No
7 overcast cool normal true Yes
8 sunny mild high false No
9 sunny cool normal false Yes
10 rain mild normal false Yes
11 sunny mild normal true Yes
12 overcast mild high true Yes
13 rain mild high true No
14 rain hot normal false
Decision Tree
8 Yes + 5 No
Outlook Root Node

Overcast

Sunny Day Temp Humidity Windy

Rain
3 hot high false
7 cool normal true
12 mild high true

3 Yes + 0 No
Terminal Node (Leaf)
Part II
Let’s do it together!
 Decision Tree
8 Yes + 5 No
Outlook Root Node

Overcast

Sunny Yes (3) Rain

Day Temp Humidity Windy

1 hot high false
2 hot high true
8 mild high false
9 cool normal false
11 mild normal true

2 Yes + 3 No
More splitting is needed
 Decision Tree
8 Yes + 5 No
Outlook Root Node

Overcast

Sunny Yes (3) Rain

Day Temp Humidity Windy

1 hot high false
2 hot high true
8 mild high false
9 cool normal false
11 mild normal true

2 Yes + 3 No
More splitting is needed
 Decision Tree
8 Yes + 5 No
Outlook Root Node

Overcast

Sunny Yes (3) Rain

High Humidity ?

no
yes

Day Humidity Windy Day Humidity Windy

1 high false
9 normal false
2 high true
11 normal true
8 high false
 Decision Tree
8 Yes + 5 No
Outlook Root Node

Overcast

Sunny Yes (3) Rain

High Humidity ?

yes no

No Yes
(3) (2)
Part III
Students in pairs
(4 mins)
 Decision Tree
8 Yes + 5 No
Outlook Root Node

Overcast

Sunny Yes (3) Rain

High Humidity ? Day Temp Humidity Windy

4 mild high false
yes no
5 cool normal false
No Yes 6 cool normal true
(3) (2) 10 mild normal false
13 mild high true

3 Yes + 2 No
More splitting is needed
 Decision Tree
8 Yes + 5 No
Outlook Root Node

Overcast

Sunny Yes (3) Rain

High Humidity ? Day Temp Humidity Windy

4 mild high false
yes no
5 cool normal false
No Yes 6 cool normal true
(3) (2) 10 mild normal false
13 mild high true

3 Yes + 2 No
More splitting is needed
 Learned Decision Tree
Outlook Root Node

Overcast

Sunny Yes (3) Rain

High Humidity ?
Windy ?
yes no
yes no
No Yes
No Yes
(3) (2)
(2) (3)
Terminal Nodes (Leaf)
 Prediction?
Outlook Root Node

Overcast

Sunny Yes (3) Rain

High Humidity ?
Windy ?
yes no
yes no
No Yes
No Yes
(3) (2) (2) (3)
Day Outlook Temp Humidity Windy
14 rain hot normal false
 Prediction?
Outlook Root Node

Overcast

Sunny Yes (3) Rain

High Humidity ?
Windy ?
yes no
yes no
No Yes
No Yes
(3) (2) (2) (3)
Day Outlook Temp Humidity Windy
14 rain hot normal false
 Prediction?
Outlook Root Node

Overcast

Sunny Yes (3) Rain

High Humidity ?
Windy ?
yes no
yes no
No Yes
No Yes
(3) (2) (2) (3)
Day Outlook Temp Humidity Windy
14 rain hot normal false
 Prediction?
Outlook

Overcast

Sunny Yes (3) Rain

High Humidity ?
Windy ?
yes no
yes no
No Yes No Yes
(3) (2) (2) (3)
Day Outlook Temp Humidity Windy
14 rain hot normal false
Part IV
Review
Predict if Peter will Play Golf?
Day Outlook Temp Humidity Windy Play
1 sunny hot high false No
2 sunny hot high true No
3 overcast hot high false Yes
4 rain mild high false Yes
5 rain cool normal false Yes
6 rain cool normal true No
7 overcast cool normal true Yes
8 sunny mild high false No
9 sunny cool normal false Yes
10 rain mild normal false Yes
11 sunny mild normal true Yes
12 overcast mild high true Yes
13 rain mild high true No
14 rain hot normal false
Predict if Peter will Play Golf?
Day Outlook Temp Humidity Windy Play
1 sunny hot high false No
2 sunny hot high true No
3 overcast hot high false Yes
4 rain mild high false Yes
5 rain cool normal false Yes
6 rain cool normal true No
7 overcast cool normal true Yes
8 sunny mild high false No
9 sunny cool normal false Yes
10 rain mild normal false Yes
11 sunny mild normal true Yes
12 overcast mild high true Yes
13 rain mild high true No
14 rain hot normal false
Splitting at the Nodes

• Entropy (information gain)

• Gini Impurity
 Learned Decision Tree
Outlook Root Node

Overcast

Sunny Yes (3) Rain

High Humidity ?
Windy ?
yes no
yes no
No Yes
No Yes
(3) (2)
(2) (3)
Terminal Nodes (Leaf)
Assumptions

• It is assumed that the attributes outside the

ones from the given table has no effect during
the life of the decision tree
Q&A