Namsor's Split Name feature relies on predictive AI models capable of correctly splitting a full name into given name and family name, regardless of the order or cultural structure of the name. The model has learned the patterns specific to each onomastic tradition from billions of names, and adapts the split to the detected cultural context, without applying hard-coded rules.
Here are some examples that illustrate the diversity of conventions correctly handled.
East Asian names: reversed order
In Chinese, Japanese and Korean, the family name comes first. Namsor recognizes this convention and reverses the order:
- 毛泽东 (Chinese) → 泽东 (given name) + 毛 (family name)
- 山田太郎 (Japanese) → 太郎 (given name) + 山田 (family name)
Arabic patronymic names
Arabic naming conventions use patronymic markers like "bin" (son of) or "bint" (daughter of). Namsor identifies the structure and places each element correctly:
- محمد بن سلمان → محمد (given name) + بن سلمان (patronymic family)
Icelandic patronyms
Icelandic names use patronyms (son/daughter of) rather than hereditary family names:
- Björk Guðmundsdóttir → Björk (given name) + Guðmundsdóttir (patronym)
Hispanic compound names
Hispanic naming conventions combine a compound given name with a double family name (paternal + maternal). Namsor preserves both:
- Gabriel García Márquez → Gabriel (given name) + García Márquez (double family name)
- María del Carmen López García → María del Carmen (compound given name) + López García (double family name)
Native scripts, no transliteration required
The feature handles names in their native writing systems without requiring transliteration to Latin first. Whether the input is in Han, Arabic, Cyrillic, Devanagari or any of Namsor's 22 supported scripts, the model splits the name directly.
These examples are only a glimpse: the model relies on patterns learned at scale and continues to perform well on rare or mixed conventions it has never explicitly encountered.