Fuzzy name matching addresses the challenges of identifying name variants within and across languages.  The winner of the MITRE Multicultural Name Matching Challenge,  NetOwl offers highly accurate, fast, and scalable name matching using an advanced machine learning-based approach.
Multi-ethnicity Name Matching Challenges

Multi-ethnicity Name Matching Challenges

NetOwl supports a wide variety of fuzzy name matching challenges including:

  • multiple transliteration variants of foreign names (Abdel Fattah el-Sisi – Abdul Fatah al-Sisi)
  • nicknames (William – Bill – Billy, Mikhail – Misha)
  • initials (John Fitzgerald Kennedy – J. F. Kennedy)
  • misspellings (Barack Obama  –  Barak Obama)
  • orthographic variations (Joaquín Guzmán –  Joaquin Guzman)
  • tokenization variations (Moon Jae-in – Moon Jae In – Moon Jaein)
  • transliteration standard differences (Xi Jinping –  Hsi Chin-p’ing)
  • optional name tokens (Joaquín Archivaldo Guzmán Loera – Joaquín Guzmán)
  • name order variations (Park Sol Mi – Sol Mi Park)
  • ethnicity-specific variations (Abd al-Aziz  –  Abdul Aziz)
  • acronyms (Bank of America – BofA)
  • alternate names (Mumbai – Bombay)
  • names in different languages (Xi Jinping – 习近平 – Си Цзиньпин – شي جين بينغ )
Revolutionary Machine Learning-based Approach

Revolutionary Machine Learning-based Approach

Traditional name matching approaches, such as Soundex, edit distance, and rule-based methods, suffer from both precision (false positives) and recall (false negative) problems in addressing the variety of fuzzy name matching challenges discussed above.

NetOwl applies an empirically driven, machine learning-based probabilistic approach to name matching challenges. It derives intelligent, probabilistic name matching rules automatically from large-scale, real-world, multi-ethnicity name variant data.

NetOwl utilizes different matching models optimized for each of the entity types (e.g., person, organization, place) In addition, NetOwl performs automatic name ethnicity detection to apply the most appropriate models to names based on their name ethnicity values in order to achieve state-of-the-art accuracy.

Multilingual and Cross-lingual Matching

Multilingual and Cross-lingual Matching

NetOwl supports multiple languages and also handles matching of a name in one language against names in other languages. NetOwl currently supports the following languages, and additional languages are on the way.

  • Languages written in Latin-based alphabets
  • Arabic
  • Chinese (traditional and simplified)
  • Languages written in Cyrillic-based alphabets
  • Greek
  • Hebrew
  • Japanese (kanji, hiragana, katakana)
  • Korean (Hangul)
  • Persian (Farsi and Dari)
Name Matching Applications

Name Matching Applications

The NetOwl name matching tool is used for many mission-critical applications in a variety of domains where failure to match may result in not only lawsuits, fines, or financial theft, but also, in the worst case, human casualties.  Example of applications include:

For applications that require matching names plus additional record fields such as date of birth, address, etc., NetOwl EntityMatcher is used instead.

Key Features of Name Matching

  • Accurate

    Accurate

    Achieves high accuracy as evidenced by winning the MITRE Multicultural Name Matching Challenge

  • Fast

    Fast

    Performs fuzzy name matching against tens of millions or more names with sub-second response times.

  • Scalable

    Scalable

    Highly scalable name matching software with Docker and Kubernetes support.

  • Customizable

    Customizable

    Allows parameter tuning as well as addition of custom rules and dictionaries

  • Easy Integration

    Easy Integration

    An easy-to-integrate fuzzy name matching tool with a REST API

  • Many Entity Types for Matching

    Many Entity Types for Matching

    Supports name matching for multiple entity types: person, place, organization, address, vehicle, date, email, nationality, and phone.

Contact us and discover what NetOwl can do for you!

    Full Name *

    Company Name *

    Email Address *

    Phone

    How can we help you? *