Decision Tree Node

The Decision Tree Node is commonly used across processes that are used to filter records through exclusion logic, apply a dynamic segmentation, or decide what kind of content a person will receive in a tweet, email, or personalized web experience. As the node is a generic piece of logic, it has a wide range of uses and applications.

General usage: 

• Each column is called a 'characteristic', and each row within a branch is a Condition
• Similar to the Columnar Table, this node executes from left to right and top to bottom
• The values supplied to each characteristic are controlled when the Decision Tree Node is placed into a graph
• As the Decision Tree Node executes each characteristic, branch, and condition,  the first combination that executes true will return its corresponding result, and all other rules will not be achieved. Hence, it is always best to have your strictest decisions on the top of each branch.
• A catch-all condition Else branch is provided by default - having a condition in this case is unnecessary. The evaluation passes straight through to the next condition; this is equivalent to the blank cell in the Columnar Table.

The example above shows a sample segmentation based on the Gender, Age, and Shoe Size of a customer shown in the left-most three columns. The segmentation outcome is shown in the Results column on the right. 

Adding a Decision Tree Node

Step 1: Select Create and choose the Decision Tree node in the 'Logic' section 

Step 2: The Decision Tree starts with one characteristic and one branch, as below. Click Add Characteristic at the top left-hand side of the window to add new characteristics and conditions to evaluate. This will prompt for the decision characteristic name, which is shown in the top row. 

Starting point for when new Decision Trees are createdAfter selecting 'Add Characteristic', a pop-up will ask what name the decision will be calledAfter. adding the Age characteristic, a new column will get added to the right, before the results Step 3: After adding a characteristic, Click on the add + icon placed throughout the decision tree to add branches and conditions to evaluate.

Creating a new branch

Step 4: After a new branch is created, you can select the dark boxes to put in a condition to evaluate. 

Step 5: Add results, which can be in the form of literals (hard-coded / static), a section of your schema, or a field in your schema (dynamic) 

The example above shows a sample segmentation based on the Gender, Age and Shoe Size of a customer shown in the left-most three columns. The segmentation outcome is shown in the Results column on the right. Each Characteristic has a minimum of two branches. There is always a catch-all branch called 'else' which will always be chosen if none of the preceding conditions is true.

 Step 6: For each characteristic created, a data source must be added to the list of parameters outside of the node at the graph level

To use the Decision Tree Node in a graph, it is necessary to map all of the named Characteristics to the schema, public variable, or literal values. These are configured in the usual way in the right-hand panel when the node is highlighted in the graph. 

Validation Warnings

: All the invalid notifications will be displayed after expanding the Invalid Drop-down.

See Also: 

• Business Logic Nodes
• Columnar Table Node
• Decision Tree Node
• 2D Table Node
• Conditional Node