1.7 Manage Pipelines

Information about managing the CS pipelines.

There are four ready-to-use pipelines: two pipelines are used for individual batch and real- time data and two pipelines are used for entity batch and real-time data. Within the pipelines are different matching configurations for SAN, PEP, EDD, and PRB records which can be tuned depending on your risk appetite.

Matching configuration is set using the Matching Rules widget. Each widget defines a match configuration for a source (customer) and target (watch list). The source and target can be filtered. This is how we set the different matching configurations for entities and individuals. It can also be used for setting different configurations for jurisdictions and domains.

Each ruleset can contain multiple rules and the score given to an alert is the maximum of the individual rule scores. An alert is generated only if the score is above the ruleset threshold. Within a rule configuration, source and target attributes are defined along with the match type (exact match, fuzzy match, or date match) and scoring method (Reverse Jaro Winkler, Jaro Winkler, Levenshtein). Each attribute level match has a threshold, below which the score is not considered, and a weightage. If the Must condition is set, then a match will not be created unless the score for that attribute is above the threshold.

To match the customer records with watch list data, run the IndividualScreening batch for an individual or Entity Screening batch for an entity in the Scheduler Service.