You create a matching decision tree to tell the add-on which data values to compare and how to compare them. Based on the comparison result you can specify that the add-on set a record's relationship status, or perform additional tests to ensure accurate results. A relationship status identifies records as positive or negative matches, or as suspected of being a match. The add-on creates groups of records identified as positive matches.
The add-on uses decision tree when:
Users manually run a match.
Inline matching is triggered, to inform users that there are potential matches (suspect records), or positive matches.
To create a decision tree:
Select the Decision tree tab.

As shown below, 1) drag a Data comparison test from the Toolbox and 2) drag the bottom connection point from the Start node and link it with the test.

Double-click the test to edit its functions:
The Name field sets a name for this test that displays on the tree.
Comparison function: the add-on uses the selected function to compare the fields included below. When you choose a function from the menu, you can select the preview icon to view a more detailed description of the function.
Minimum score: Sets the minimum score to consider this data comparison test a success and evaluate to true.
Fields: One or more fields used in the comparison. The Algorithm property determines which algorithm is used to compute a score for this field. When using multiple fields as part of one data comparison, the Weight property is used to compute a weighted average and assign a weight to each field.
Specify the relationship status if the test evaluates to true by dragging one of the following from the Toolbox and connecting it with the test's right output point:
Match: the records will be considered a match.
Suspect: the records are suspected of matching, but require human review.
No Match: the records do not match.

Specify the relationship status if the test evaluates to false by connecting the desired status to the test's left output point (shown in left-hand image). Note that if you want to have the evaluated output trigger a new test, add another Data comparison and connect it to desired output (shown in right-hand image).

Use the process in the previous steps to build out your matching tree. Once finished, select Save and close.
What's next?Task 4: Configuring a merge policy gives instructions on setting up the add-on to merge data.