farbodbij/persian-gender-by-name

Persian Gender Detection by Name

A comprehensive dataset for determining gender based on Persian names, enriched with English representations.

Overview

The Persian Gender Detection by Name dataset is the largest of its kind, comprising approximately 27,000 entries. Each entry includes a Persian name, its corresponding gender, and the English transliteration. This dataset is designed to facilitate accurate gender detection and enhance searchability through multiple name representations.

Features

• —Extensive Data: ~27,000 name-gender-English tuples.
• —Multiple Representations: Various spellings and formats for each name to improve search flexibility.
• —High Quality: Aggregated from reliable sources and meticulously hand-cleaned for accuracy.
• —Expandable: Plans to incorporate more names and data sources in the future.

Data Sources

This dataset aggregates information from the following primary sources:

• —Iranian Names Database By Gender
• —Persian Names Gender Dataset on Kaggle
• —Persian Names with Gender and Transliteration Data

Additionally, supplementary data was scraped and manually cleaned to ensure consistency and completeness.

Data Structure

The dataset is organized in a CSV format with the following columns:

• —Name: The Persian name.
• —Gender: Assigned gender (e.g., Male, Female).
• —English Representation: The transliterated version of the Persian name.

Example:

Usage

This dataset is ideal for:

• —Developing gender prediction models based on Persian names.
• —Academic research in linguistics, gender studies, and natural language processing.
• —Enhancing search algorithms with multilingual name representations.

Future Enhancements

Future updates will focus on:

• —Expanding the dataset with additional names and gender associations.
• —Incorporating more diverse sources to cover a broader range of names.
• —Refining data quality through ongoing cleaning and validation processes.

Citaion

To use please cite the following paper:

@misc{bijary2025agenticusernamesuggestionmultimodal,
      title={Agentic Username Suggestion and Multimodal Gender Detection in Online Platforms: Introducing the PNGT-26K Dataset}, 
      author={Farbod Bijary and Mohsen Ebadpour and Amirhosein Tajbakhsh},
      year={2025},
      eprint={2509.11136},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2509.11136}, 
}

Acknowledgments

Thanks to the contributors of the original datasets and those who assisted in data aggregation and cleaning.