Semantic entity and attribute matching.
Understanding the meanings behind the strings and numbers that make up the data record is key for precise matching, to avoid false positives (similarly looking but distinct values) and false negatives (missing good matches).
Step 1 is parsing the data to understand it, splitting it up into its components. Each matcher makes use of its corresponding Parser internally.
Type-specific dictionaries and matching logic is applied, and complex result objects are returned to the API caller.
Matchers are available on every hierarchy level, like a system of Lego building blocks. For simple 1-value attributes such as a Date, to entities like a postal address. Either as plain text values in a string, or semi-structured, to fully split into designated fields.
The matcher stack is an integral part of our SearchCluster turn-key solution.