What’s New: Customizing the Name Matcher to Your Data
We’re excited to announce the release of the latest version of Rosette Name Translator and Rosette Name Indexer.
Highlights of the Release include:
- ENHANCED: Support for indexing, querying and matching Russian names
- Multi-threaded support for adding/deleting names from the name matching index
- Easier integration of name matching with Solr (your package now comes with the jar files that you can drop into the proper directories to access Rosette instantly)
- Web services API for accessing name translation functionality—with name matching coming soon in a future release.
- Greater control over name matching results (to override default matching behavior)
This last feature is one that many of our customers have been asking for both to better fit the matcher to their data and increase accuracy of matches for the various applications using Rosette Name Indexer.
We recognize that determining if a name match is good or not can be heavily dependent on the user’s data, so we’ve added three properties which give users better control over the match confidence score.
First, stop patterns allow users to designate words that should be ignored during the matching process, such as titles “Sir,” “Mayor,” “Dr.” etc. This feature is available to all supported languages.
Second, if a user has two relatively unrelated full names, but would like them to be designated a match, that is now possible. This feature is especially useful in the case of aliases—such as the American-turned Taliban fighter, “John Walker Lindh” also known as “Abu Sulayman al-Irlandi,” which are two names that would normally never come up with a match score. Conversely, two names that Rosette Name Indexer would normally designate as a match can also be set to “not match.” This feature is available for English, Japanese, Chinese, and Russian.
The third case is similar to the one above, but works on single name components, such as “Cyndi” and “Cindy,” instead of a full name “Cindy Smith” and “Cyndi Smyth.” This control serves to boost the match of name pairs such as “Abigail Smith” and “Abby Smith” by the user indicating that “Abigail” and “Abby” are a strong match. This version of Rosette Name Indexer provides a list of English names and nicknames, which can be edited by the user.