Scribd
Upload
15 views3 pages

Decision Trees Report

Uploaded by

manon.mustafa.325
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
15 views3 pages

Decision Trees Report

Uploaded by

manon.mustafa.325
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Decision Trees:

Are a supervised learning method used for classification and regression.


- A Decision Tree is a flowchart-like tree structure where:
 Internal node represents features
 the branch represents a decision rule .
 Each leaf node represents the outcome .

Assumptions while creating Decision Tree

1-In the beginning, the whole training set is considered as the root.

2-Feature values are preferred to be categorical. If the values are continuous, then they are
discretized prior to building the model.

3-Records are distributed recursively on the basis of attribute values.

4-The order of placing attributes as the root or internal node of the tree is done by using
some statistical approach.
How does the Decision Tree algorithm work?

 The basic idea behind any decision tree algorithm is as follows:


o Select the best attribute using Attribute Selection Measures (ASM) to split the records.
o Make that attribute a decision node and break the dataset into smaller subsets.
o Start tree building by repeating this process recursively for each child until one of the conditions is met:
 All the tuples belong to the same attribute value.
 There are no more remaining attributes.
 There are no more instances

Pros

 Decision trees are easy to interpret and visualize.


 They can easily capture non-linear patterns.
 They require fewer data preprocessing steps from the user; for example, there is no need to normalize columns.
 They can be used for feature engineering, such as predicting missing values, making them suitable for variable
selection.
 The decision tree has no assumptions about distribution because of the non-parametric nature of the algorithm.

Cons

 Sensitive to noisy data, which can lead to overfitting.


 Small variations (or variance) in data can result in different decision trees. This can be reduced by bagging and boosting
algorithms.
 Decision trees are biased with imbalanced datasets, so it is recommended to balance the dataset before creating the
decision tree.

Binary DT VS. Multiple DT

A binary decision tree is a type of decision tree where each internal node has exactly two child nodes, which
correspond to the two possible outcomes of a binary decision.

The tree is constructed by recursively partitioning the input space into two regions based on the value of a
chosen feature at each internal node until a stopping criterion is met.

Binary decision trees are commonly used in classification tasks, where the goal is to assign a discrete label
to an input based on its features.

On the other hand, a multiple decision tree, also known as a multiclass decision tree, is a decision tree that
can handle more than two classes.

A multiple decision tree, also known as a multiclass decision tree, is a decision tree that can handle more
than two classes. In other words, each internal node can have more than two child nodes, each
corresponding to a different class or outcome.

The tree is constructed by recursively partitioning the input space into multiple regions based on the value
of a chosen feature at each internal node until a stopping criterion is met.
Multiple decision trees are commonly used in classification tasks where the goal is to assign a label to an
input based on its features.

Issues in Decision Trees


 How deeply to grow the decision tree?
 Handling continuous attributes
 Choosing an appropriate attribute selection measure
 Handling training data with missing attribute values
 Tendency to over-fit:
o If your tree goes too big (deep), you will overfit pretty fast!
o It's a tree-like structure, so you will lose some relationships between features.

 They use greedy learning algorithms:


o Because we usually use greedy traversal when dealing with a tree.
o This is a downside because it is not guaranteed to find the global optimal model.

You might also like