Configuring system matching

System matching uses system match rules to identify matching records. To configure system matching, you define and deploy your system match rules on the i2 Analyze server.

The mechanism that you use to deploy system match rules depends on your deployment pattern:

  • In a deployment that contains the Information Store, your system match rules are used to create a match index that determines whether a record matches one that exists in the Information Store. The match index is employed by the Get Matches feature in Analyst's Notebook, and when records are presented through import, connectors, or created by analysts. To create the match index from your match rules, see Deploying system match rules.

  • In a deployment that contains only the i2 Connect gateway, a match index is not necessary because the system match rules act only on records in search results and on the chart surface. To deploy your system match rules in this deployment pattern, see Deploying system match rules for the i2 Connect gateway only.

Note: If your deployment does not contain a system match rules file, Analyst's Notebook will still use source identifier matching to determine if records match each other. To change this behavior, review the instructions linked above.

Analysts can choose whether to use system matching at all, but they cannot choose the match rules that are used or approve matches on a record-by-record basis. To ensure that your system match rules do not cause false-positive matches, focus on using known strong identifiers in your data.

If system matching acts on data that involves a large number of seed or matching records (in the tens of thousands), the matching process can take a long time to complete. In this scenario, the following actions might accelerate the process:

  • Investigate your match rules to ensure that you are not matching too many records by accident.

  • Investigate that the data in your Information Store is being ingested correctly, so that your strong identifiers are not accidentally present on too many records.

  • Advise your analysts to use system matching on smaller sets of data.