Best Practices for Tuning Name Search Parameters

Name matching can be tricky — especially If you’re not familiar with its intricacies, like nicknames or transliteration errors which are common when dealing with foreign languages. For organizations and government agencies with critical name matching requirements, like screening against sanctioned party lists, errors could have serious consequences.

Standard search engines fall short when it comes to detecting subtle differences between names or accounting for non-Latin characters. In response, many organizations have turned to name matching systems based on statistical models that provide the foundation for machine learning.

While those models are generally quite good out-of-the-box, they can become even better with some adjustments to their configurations. In this webinar, Michael Harris Solutions Engineer and Pat Deeb, Senior Solutions Engineer from Babel Street will explore some best practices around tuning your name matching system for the best results according to your organization’s needs.

In this webinar you will learn:

• Considerations for setting thresholds around “what is a match”
• Setting scoring penalties for names that aren’t in order or that may have middle initials
• Using entity types to account for person names and organization names
• The best way to input addresses

They also demonstrated how to easily accomplish search tuning with Rosette Match Studio – a user-friendly interface to Rosette Name Indexer.