An artificial intelligence to estimate the gender of a name

Artificial intelligence and morphology of a name

What is the morphology of a name?

Morphology deals with the morphemes that exist in a language and how words are constructed. Morphemes are the smallest units of a language that have some kind of grammatical meaning. These are things like roots (book, man, tree) and affixes (un-, -s, -ly, dis-).
The type of a personal name and the language system to which these names adhere to are two very significant components to determine the morphological pattern of a personal name.

Why use artificial intelligence?

Thanks to years of research, our many contributions to the world of academic and scientific research, and our work for international and regional organizations, we benefit from one of the most comprehensive databases in the world.

Databases are not always error-free

However, a simple database search would give accurate results, albeit with a significant margin of error. That's why we've trained an artificial intelligence system to detect errors in the databases and in the data you send it: such as typos or reversals of first and last names.

A database does not contain all existing data

No matter how complete a database may be, not all the first and last names in the world will be included. This is particularly the case for nicknames, or for very recent and very uncommon names. A simple database search using a standard tool will normalize the name, generating a very high error rate.

Our tool is capable of analyzing the morphology of even the most unique name in the world to determine its gender and origin with the utmost precision.

A database search doesn't take the local context into account

What's more, the gender of a name differs from one region of the world to another, and even within the same region but from a different community. This is why Gender Guesser uses Namsor's technology to determine the local context of the name, i.e. its origin and possible diaspora. By taking into account the local context or cultural context, the system can analyze the name in relation to its surroundings, historical influences, and contemporary trends, which enhances its ability to estimate the gender associated with that name more accurately.

Gender Guesser under the hood

How our gender estimation technology is designed:

1. Get data: Retrieval of data, mostly lists of names correlated to genders or to a geographical context.
2. Prepare data: Data preparation, data quality check, sorting, classification, cleaning.
3. Train model: Training of an artificial intelligence specially adapted to onomastics with the prepared datasets.
4. Test data: Testing the ability of artificial intelligence to correctly predict information from proper names.
5. Continuous improvement: Added new feature with continuous improvement process.