The Matching Rules widget enables you to define the matching configuration for a set of data. This widget should be configured only if watchlist screening for Primary/Secondary Prospects is to be performed as part of KYC Risk Assessment.
The data that must be matched by each widget depends on the source and target set in the entity widgets and which link to the matching rules widget. The source and target data can be filtered if a subset of data is to have this matching configuration applied. This allows you to provide different matching configurations for different types of watchlist records and different jurisdictions and domains. Each matching ruleset consists of name, description, scoring aggregation used, the threshold value for the overall rule set and one or more rules.
Rules are configured using the Matching Ruleset window. Matches are generated based on a defined set of attributes for each rule. A weighted average of the score is generated for each of the attribute level matches. Provide the following values for each rule:
· Source attribute
· Target attribute
· Match type
· Scoring method. This can be one of the following:
§ Levenshtein: The Levenshtein Distance (LD) or edit distance provides the distance, or the number of edits (deletions, insertions, or substitutions) needed to transform the source string into the target string. For example, if the source string is Mohamed and the target string is Mohammed, then the LD = 1, because there is one edit (insertion) required to match the source and target strings. Click here for more information.
§ Jaro Winkler: The Jaro Winkler similarity is the measure of the edit distance between two strings. For example, if the source string is Mohammed Ali and the target string is Mohammed Ali, then the similarity = 1. Click here for more information.
· Set threshold value which if crossed means that the attribute is considered for matching
· Weightage assigned to the attribute (total of all attributes within a rule must equal 1)
· Must check box (optional). If this check box is selected, then there must be a match on this attribute; if not, no matches are generated for this rule.
Each combination of attributes in the match rule will be scored. If the threshold for an attribute is greater than the specified attribute level threshold then the score contributes to the overall score. If data is null for either the source or target attribute a score of 50 is given. Attribute level scores are multiplied by the weightage and then added to get the weighted average score for the customer and watchlist record. If the score is greater than the rule threshold, then the record is considered for matching.
If there are two or more rules in the ruleset then the maximum score is taken. If this score is greater than the threshold defined for the ruleset, than the two records are a match.
To add a ruleset, follow these steps:
1. In the Pipeline Designer page, select the pipeline you want to edit. The Pipeline Designer window appears.
2. Hover on the Matching Rules widget and
click Edit . The Matching
Ruleset window is displayed.
3. Select a prospect type.
4. Provide details as described in the following table:
Matching Rules Widgets and their Descriptions
Field |
Description |
Ruleset Name |
Enter the name for your ruleset. This is a mandatory field. |
Description |
Enter the description of the ruleset. This is a mandatory field. |
Scoring Aggregation Type |
Select the scoring type. Currently, only Maximum is available. |
Set Threshold |
Enter the threshold value for the ruleset. |
Source |
Select Filter Attributes can be Business Domain Code, Customer Type Code, or Jurisdiction Code. Enter the value based on the attribute. For example, a value for jurisdiction code can be JC1. Click
Save |
Target |
Select Filter Attributes can be Entity Type, WatchList Type, or WatchList Sub Type. Enter the value based on the attribute. For example, a value for watchlist type can be SAN. Click
Save |
Rules |
Select Add |
Name |
Enter the rule name. |
Description |
Enter the description of the ruleset. This is a mandatory field. |
Rule Threshold |
Enter the threshold value for the rule. |
Mappings |
Select Add |
Source Attribute |
Select one or more source attributes from the customer record that must be matched. |
Target Attribute |
Select one or more attributes from the watch list against which matching is performed. |
Match Type |
Select the matching type. The following match types are available: · Exact · Fuzzy · Date |
Scoring Method |
Select the scoring method if you have selected the match type as Fuzzy. The following scoring methods are available: · Levenshtein · Jaro Winkler · Reverse Jaro Winkler |
Threshold |
Enter the threshold score. |
Weightage |
Enter the weightage. |
Condition |
If this check box is selected, then this condition must be met for matching. |
5. Click Save to save the changes.
The rule is created and is visible on the canvas. It is also available
for use in the Matching
Ruleset window.
To return to the Pipeline Designer
window, click Close to
close the window. Finally, click Save
to save the updates made.