0% found this document useful (0 votes)
5 views32 pages

CART Decision Tree: Overview & Techniques

The document discusses the CART decision tree algorithm, which can handle both classification and regression tasks using the Gini index for decision points. It highlights the limitations of the ID3 algorithm, the importance of hyperparameter tuning to prevent overfitting, and methods such as pruning to improve decision tree performance. Additionally, it introduces the concept of Random Forests, which applies decision trees to subsets of data for better accuracy.

Uploaded by

rkr201759
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
5 views32 pages

CART Decision Tree: Overview & Techniques

The document discusses the CART decision tree algorithm, which can handle both classification and regression tasks using the Gini index for decision points. It highlights the limitations of the ID3 algorithm, the importance of hyperparameter tuning to prevent overfitting, and methods such as pruning to improve decision tree performance. Additionally, it introduces the concept of Random Forests, which applies decision trees to subsets of data for better accuracy.

Uploaded by

rkr201759
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd

Artificial Intelligence and Machine

Learning

Decision Tree – CART


Dr. Rajashree Nayak
DECISION TREE – CART
 CART is an alternative decision tree building
algorithm which can handle both
classification and regression tasks.

 This algorithm uses a new metric named gini


index to create decision points for
classification tasks.

 Select the feature that has lower Gini index


for splitting.
DECISION TREE – CART
DECISION TREE – CART
DECISION TREE – CART
DECISION TREE – CART
DECISION TREE – CART
DECISION TREE – CART
OBSERVATION
 ID3 is the most common conventional
decision tree algorithm but it has
bottlenecks. Attributes must be nominal
values, dataset must not include
missing data, and finally the algorithm tend
to fall into overfitting.
WHEN TO STOP SPLITTING?

 Usually, real-world datasets have a large


number of features, which will result in a
large number of splits, which in turn gives
a huge tree.

 Such trees take time to build and can lead


to overfitting. That means the tree will give
very good accuracy on the training dataset
but will give bad accuracy in test data.
HYPERPARAMETER TUNING
 There are many ways to tackle this problem through
hyperparameter tuning.

 We can set the maximum depth of our decision tree


using the max_depth parameter. The more the value
of max_depth, the more complex your tree will
be.

 Another way is to set the minimum number of


samples for each spilt. It is denoted
by min_samples_split. Here we specify the minimum
number of samples required to do a spilt.

 That means if a node has less than 10 samples


then using this parameter, we can stop the further
splitting of this node and make it a leaf node.
DECISION TREE – PROS AND CONS
OVERFITTING
 A hypothesis h is said to overfit the training data if there is
another hypothesis h’, such that h has a smaller error than
h’ on the training data but h has larger error on the test data
than h’.

accuracy
On training

On testing

Complexity of tree
OVERFITTING
Outlook
 Outlook = Sunny,
 Temp = Hot
 Humidity = Normal Sunny Overcast Rain
 Wind = Strong 1,2,8,9,113,7,12,13 4,5,6,10,14
 label: NO 2+,3- 4+,0- 3+,2-
 this example doesn’t exist in the Humidity Yes Wind
tree

High Normal Strong Weak


No Yes No Yes
OVERFITTING
This can always be
Outlook
done – may fit noise
or other
coincidental
• Outlook = Sunny, regularities Rain
Sunny Overcast
• Temp = Hot
1,2,8,9,113,7,12,13 4,5,6,10,14
• Humidity = Normal
2+,3- 4+,0- 3+,2-
• Wind = Strong
Humidity Yes Wind
• label: NO
• this example doesn’t exist in the
tree High Normal Strong Weak
No Wind No Yes

Strong Weak
No Yes
REASONS FOR OVERFITTING
 Too much variance in the training data
 Training data is not a representative sample
of the instance space
 We split on features that are actually irrelevant

 Too much noise in the training data


 Noise = some feature values or class labels are
incorrect
 We learn to predict the noise

 In both cases, it is a result of our will to minimize the


empirical error when we learn, and the ability to do it
(with DTs)
PREVENTING OVERFITTING
PRUNING

 Pruning is another method that can help us avoid


overfitting. It helps in improving the performance
of the Decision tree by cutting the nodes or
sub-nodes which are not significant.
Additionally, it removes the branches which
have very low importance.
 There are mainly 2 ways for pruning:
 Pre-pruning – we can stop growing the tree
earlier, which means we can prune/remove/cut a
node if it has low importance while growing the
tree.
 Post-pruning – once our tree is built to its
depth, we can start pruning the nodes based on
their significance.
DECISION TREE TO RANDOM FOREST

 Instead of applying decision tree algorithm on all dataset,


dataset would be separated into subsets and decision tree
algorithm will be applied to these subsets

 Decision would be made by the highest number of subset


results
GRAPHICAL ABSTRACT
HOW DOES IT WORK ?
HOW DOES IT WORK ?
HOW DOES IT WORK ?
HOW DOES IT WORK ?
HOW DOES IT WORK ?
ADVANTAGES

You might also like