Adaboost
Description	Adaboost is a boosting algorithm that combines multiple weak models into a single strong learner algorithm (the predictive model).
Why to use	To classify text into the possible categories.
When to use	When textual data needs to be classified.	When not to use	When the dataset does not contain textual data.
Prerequisites	The dataset should not contain any missing (NaN) values. The dataset should contain at least one categorical and one textual variable.
Input	Textual Data	Output	Classified data with predicted labels
Statistical Methods used	Accuracy Sensitivity Specificity F-score Confusion Matrix	Limitations	Weak classifiers (being too weak) may lead to overfitting Adaboost is sensitive to outliers

Adaboost is located under Textual Analysis ( ) in Classification, in the left task pane. Use the drag-and-drop method to use the algorithm in the canvas. Click the algorithm to view and select different properties for analysis.

Refer to Properties of Adaboost.

Properties of Adaboost

The available properties of Adaboost are as shown in the figure below.

The table below describes different fields present on the properties of Adaboost.

Field		Description	Remark
Task Name		It is the name of the task selected on the workbook canvas.	You can click the text field to edit or modify the task's name as required.
Dependent Variable		It allows you to select the dependent variable.	You can select one variable.
Text		It allows you to select the text variable.	The default value is eight (8).
Advanced	Random State	It allows you to select a random combination of train and test for the classifier.	It randomizes the data splitting into train and test. This provides different data each time for training and testing the model. The model is expected to return the same result each time, even with different train-test combinations. It ensures that the obtained results can be reproduced. You can enter any integral value as the random state This parameter is optional.
	Learning Rate	It allows you to select the weight value for the learning rate of weak learning algorithms.	By default, the value for the learning rate is 0.1. You can enter any float value between 0 and 1. A low learning rate indicates that the contribution by each learner is less. In that case, the number of estimators should be kept high.
	Number of Estimators	It allows you to select the number of weak learners to be created while running the algorithm.	By default, the number of estimators selected is 50. You can enter any integral number of estimators. The number of estimators should be decided based on the selected Learning Rate.
	Algorithm	It allows you to select the boosting algorithms for classification. There are two types of boosting algorithms. · Discrete boosting algorithm · Real boosting algorithm	Algorithm selection is important, especially when you have multi-class datasets. You can select any one of the following algorithms: SAMME (discrete) SAMME.R (real) SAMME considers the discrete values while SAMME.R considers the probability values. Usually, SAMME.R converges faster and creates low test errors with few boosting iterations. By default, the SAMME.R algorithm is used.
	Add result as a variable	It allows you to select any of the result parameters as the variable.	You can select from the following performance parameters of SVM: Accuracy Sensitivity Specificity F-score
Node Configuration		It allows you to select the instance of the AWS server to provide control on the execution of a task in a workbook or workflow.	For more details, refer to Worker Node Configuration.
Hyperparameter Optimization		It allows you to select parameters for optimization.	For more details, refer to Hyperparameter Optimization.

Example of Adaboost

Consider a dataset containing BBC news snippets classified into different categories like tech, business, sport, politics, and so on.

A snippet of input data is shown in the figure given below.

We select the following properties and apply Adaboost.

Dependent Variable	*Category*
Text	Text
Random State	None
Learning Rate	0.1
Number of Estimators	5
Algorithm	SAMME

The result page is displayed in the figure given below. On the result page, you can see the following output parameters:

Accuracy, F-score, Precision, Sensitivity, and Specificity are the model performance indicators displayed on the Results page.
The Accuracy of the above Adaboost model is 0.7368, F-score is 0.7475, Precision is 0.7169, Sensitivity is 0.8143, and Specificity is 0.8814

On the same result page, you see the confusion matrix containing predicted and actual values for Category.

The shaded diagonal cells show the correctly predicted categories. For example, 119 news (out of 140) in the business category are correctly predicted.
The remaining cells indicate the wrongly predicted categories. For example, 3 news in the business category is wrongly predicted as the tech

The figure below shows the resultant dataset in the Data tab.

Along with the Index and Text columns, you can see the Category column showing the original categories.
The predicted categories for each text are listed under the Label
You can compare Label and Category columns to check correctly and wrongly predicted categories.

Table of Contents