Fuzzy Name Matching Helps to Verify E-Signatures

E-Signatures Are a Critical Part of Business Today

In many instances, an e-signature company may want to offer additional reassurance to its customers that the name of the person who signs a document matches the name of the person who was expected to sign the document.

The process adopted may vary in its specifics, but it generally follows these steps:

• A sender requests an e-signature from a person (receiver)
• The receiver e-signs the document
• The sender receives back a version of the document with the e-signature added
• The e-signature is verified by comparing the signed name against the expected name of the receiver to ensure that the signature is valid.

In addition, the e-signature company needs to implement controls to allow the end-user to tighten or loosen the acceptance criteria (i.e., the required level of matching) depending on the use case.

Names Can Vary Enormously for Many Reasons

A major challenge in verifying an e-signature is that names can legitimately differ for a whole host of reasons. Here are just a few examples:

• Simple misspellings and spelling variations, including names that sound alike: Sean vs. Shawn vs. Shaun
• Initials: John Fitzgerald Kennedy vs. J.F. Kennedy
• Missing name elements: John Robert Smith vs. John Smith
• Nicknames: John vs. Johnny; Mikhail vs. Misha
• Name Order Variants: Park Jae-in vs. Jae-in Park. (Asian names traditionally place the surname, e.g., Park, first, but they occasionally occur in the Western order.)
• Transliteration variants: Abdel Fattah el-Masri vs. Abdul Fatah al-Masri. (A language like Arabic is written in a script different from Latin and also has some sounds that don’t occur in English. When transforming the name from Arabic letters to English ones, differences in spelling frequently arise.)

For additional examples of types of name variation, see here.

Matching Names Across Different Writing Systems Is a Particular Challenge

It’s bad enough if the names in e-signatures only come in Latin script but they may also come in a non-Latin script that has to be authenticated against the expected Latin version.

• For example, the e-signature created by the sender may be Kim Ji-Soo, but the recipient signs it 김지수.
• Or a signature Leonid Kogan may be met by Леонид Коган.
• Or Usama ibn Munqidh vs. أسامة بن منقذ

In all these cases, the names refer to the same person but are in different writing systems.

These are some of the challenges of matching names cross-script:

• Some alphabetic languages, particularly Arabic, do not write out the vowels for the most part. For example, the name Ahmed bin Salman is written in Arabic as احمد بن سلمان. Spelled out in English letter for letter, the name is ‘hmd bn slman (‘ is a symbol for a special kind of guttural sound).
• Hebrew does not have the sound represented by English th. It will substitute a letter that is somewhat similar to it, e.g., John Douthat will be transliterated as ג’ון דוטה. The letter ט represents a sound that is only an approximation for English th.
• In many cases, the sound structure of one language will be enormously different from another, resulting in large differences in transliteration. For example, Donald Trump transliterates into Korean as 도날드 트럼프, which can be transliterated as donaldeu teuleompeu

For more information on the challenges of cross-language name matching, see here.

Advanced Fuzzy Name Matching Can Support E-Signature Verification

There’s an advanced Fuzzy Name Matching technology that offers a way to match all of these variants and more.

Advanced Fuzzy Name Matching uses intelligent machine learning algorithms to automatically learn a collection of probabilistic name matching rules from large-scale, real-world name variant data. Since the rules are learned from real data, they are not bound by limitation of humans’ knowledge of what are possible name matches. They reflect countless name variants that occur in the real world.

In the case of cross-script name matching, the matching is performed directly, with no need to translate non-Latin scripts into Latin prior to matching. This avoids the problem of mistakes introduced by the translation process. Matching directly between the names in different scripts achieves higher accuracy.

Advanced Fuzzy Name Matching allows any organization or person that authenticates documents to effectively match and verify the name in an e-signature against the expected one.