Configuring matching

In a deployment of i2 Analyze that contains data that originates from multiple sources, the likelihood of multiple records that represent the same real-world object or relationship increases. To enable analysts to identify and reduce the number of matching records, you can activate the matching features in Analyst's Notebook by creating custom match rules.

i2 Analyze can be presented with data from multiple sources, through multiple routes. The mechanisms that identify matching records depend on how the data is presented to i2 Analyze:

  • For records that are on a chart, system matching can compare them with records in the Information Store, and Find Matching Records can compare records on the chart with each other.

  • When data is presented through Analyst's Notebook import, or retrieved by connectors, or created by analysts, system matching can compare incoming records with records on the chart and in the Information Store.

  • During the ingestion process, the incoming records are compared with the records in the Information Store. For more information about identifying matches in this way, see Information Store data correlation.

Important: The scope of match rules is not the same for link records as it is for entity records. Rules for matching entity records apply universally, but rules for matching link records apply only to links between the same pair of ends. Link records that do not connect the same ends are never a match for each other.

This behavior also means that rules for matching link records behave differently in system matching and Find Matching Records:

  • In system matching, link records can match only when they are between the same pair of entity records.

  • In Find Matching Records, link records can match when they are between any two records in a pair of entity items on the chart surface.

If the pair of entity items on the chart contain a large number of united records, some combinations of those records are not searched for matching links. By default, the number of link end record combinations that are searched is 100. You can change this value, but the link matching process might take longer if more combinations are searched. For more information about changing the maximum number of combinations that are searched, see The DiscoServerSettingsCommon.properties file.