What is fuzzy logic?

Both pdNickname 2.x and pdGender 2.x are fully compatible with fuzzy logic. In these products, fuzzy logic involves slight variations in first names and nicknames based on common typographical errors and stylized spelling methods. The Pro edition of these packages comes equipped with fuzzy logic out of the box. Fuzzy logic add-ons can be appended to both the Pro and Standard versions.

The following illustrates the fuzzy logic technology employed in pdNickname 2.x and pdGender 2.x. Further information specific for these packages can be reviewed in the product user documentation found on our support page.

Typographical errors

A large majority of fuzzy logic records involve common typographical errors. These algorithms look at frequently reversed digraphs (a pair of letters used to make one phoneme or distinct sound), phonetically transcribed digraphs, double letters typed as single letters, single letters that are doubled, and other common data entry issues. The most likely typographical errors are determined based on the number of letters, the characters involved, where they are located in the name, and other factors.

The following are examples of fuzzy logic based on common typographical errors:

Example 1 | Real: AL | Fuzzy: ALL | the “L” is repeated
Example 2 | Real: ROCCO | Fuzzy: ROCO | the second “C” is left out
Example 3 | Real: CHRISTOPHER | Fuzzy: CHRISTOFER | the “PH” digraph is phonetically transcribed as “F”
Example 4 | Real: SOPHIA | Fuzzy: SOHPIA | the “PH” digraph is reversed
Example 5 | Real: MARGARET | Fuzzy: MARGRAET | the second “AR” digraph is reversed

Stylized spellings

Other fuzzy logic records involve stylized spelling methods. These algorithms look at non-regular characters such as extended ANSI characters (ASCII values 128 to 255) as well as hyphens, apostrophes, and spaces.

A few of the possible extended characters are “Á” (A-acute), “Ö” (O-umlaut), and “Ñ” (N-tilde). In these cases, “Á” becomes “A” (A-regular), “Ö” becomes “O” (O-regular), “Ñ” becomes “N” (N-regular), and other extended characters are treated similarly.

The following are examples of fuzzy logic based on stylized spellings:

Example 6 | Real: BJÖRK | Fuzzy: BJORK | spelled with O-regular instead of O-umlaut
Example 7 | Real: NICOLÁS | Fuzzy: NICOLAS | spelled with A-regular instead of A-acute
Example 8 | Real: ‘ASHTORET | Fuzzy: ASHTORET | spelled without an apostrophe prefix
Example 9 | Real: ABD-AL-HAMID | Fuzzy: ABDALHAMID | spelled without hyphens delimiting the name parts
Example 10 | Real: JUAN MARÍA | Fuzzy: JUANMARIA | spelled without the space between the two parts and with I-regular instead of I-acute

Fuzzy logic add-on packs and upgrades

Peacock Data releases additional fuzzy logic records nearly every month for pdNickname 2.x and pdGender 2.x in the form of add-on packs which can easily and economically be appended to the main databases extending coverage of typographical errors and stylized spelling methods.

The fuzzy logic technology built into the main Pro product downloads is designed to pick up statistically the most likely mistakes and stylizations. Fuzzy Logic Add-on Packs are designed to pick up less common mistakes and stylizations.

Add-on packs include new algorithms and randomizers and are fully compatible with both the Pro and Standard editions of these packages.

Those licensing the Standard edition of either product can also purchase a Standard to Pro Upgrade Pack which includes all the fuzzy logic records from the Pro edition. Once a Standard version is upgraded, it will be the same as the Pro edition.

Review the documentation provided with the fuzzy logic add-on packs and upgrades for further instructions.

Leave a Reply