Fast Accurate Match Extractor

“Innovative Cross Breed Name Matching Algorithm”

Last update

* 08/16/2012

The Name Matching Algorithm

Name matching is the core problem of an increasing number of applications (CRM, Name Vetting, De-duplication etc...). There are many available algorithms such as token based (Qgram, DiceSimilarity ...), character based (Edit Distance, levenstein, Jaro...) and phonetic based (DoubleMetaphone, Soundex ...); each provides a unique feature but exposes a weakness. FAME was born by joining the characteristic of most name matching algorithms, fixing up the weakness and amplifying the power of each, without compromising the speed. A sample csv with 1 million records loads in 15 minutes and scanning a name against the file takes less than 0.3 milli seconds on a core 2 duo.

Name Matching Data Flowchart

Applications and Uses

FAME API is a very useful tool for all .net developers who are creating any of the below applications. Business users, specifically compliance officers & managers can benefit from the free search against updated list from OFAC, EULIST and HM treasury:

Vetting persons against a watch-list (name verification)
String matching & comparison
Eliminating duplicate records (de-duplication)
Customer Relationship Management (CRM)
Identity verification
Improve document search
Improve social network analysis
Ontology Enrichment

Name Variations

Phonetic variations
Missing name segments
Additional name segments
Transliteration spelling differences
Initials
Typing mistakes
Out of order name segments
Variable segmentation
Name prefix/suffix
Titles and honorifics
Truncations

Name Variations (Company Name Matching Algorithm)

Company name variations can also be a hassle for name matching where each culture has specific abbreviation, business type, abbreviated name etc... FAME provides the algorithm to match these variations. E.g.: "Equipment Master Limited" vs. "Equipment Master Ltd".

Scoring Factors

FAME API calculates the similarity based on many factors, and returns a score between 0 and 1 with three decimal point precision (Future releases will enable users to configure these factors).

Segment misplacement	"Donna Doe" vs. "Doe Donna"
Character misplacement	"Frank Foe" vs. "Frnak Foe"
Phonetic pronunciation	"Jane Doe" vs. "Jain Doe"
Initials	"Robert Roe" vs. "Robert R."
Missing segment	"Larry Loe" vs. "Larry L. Loe"

References & Detailed Analysis

For more detailed comparison between available algorithms and how FAME solves the name matchign problem refer to our white paper.

Free Search Source Details

FAME demo gives you a free search tool against updated lists from:

The Specially Designated Nationals (SDN) List, maintained by the Office of Foreign Assets Control (OFAC) in the US Treasury Department.
The consolidated EUlist containing the names and identification details of all persons, groups and entities targeted by financial sanctions, maintained and collated by European Union.
The consolidated list of asset freeze targets designated by the United Nations, European Union and United Kingdom under legislation relating to current financial sanctions