Comparative AI ‐Based Framework for Name‐Based Demographic Inference in the Indian Subcontinent
Engineering Reports · 2026-06 · 8 (6)
DOI: 10.1002/eng2.70902 ↗Kurz gefasst
Fachartikel in der Zeitschrift Engineering Reports, online erschienen am 16.06.2026. Verfasst von Sherin Sultana und 5 weiteren Autor:innen. Die Arbeit nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
In the era of modern technology, where most activities rely on digital systems, personal information is a central concern. However, due to a lack of accurate personal data, such as correct names, nationalities, religions, and genders, they are often restricted from accessing essential technological and healthcare services. In this context, researchers have compiled a well‐structured dataset that includes the names, nationalities, religions, and genders of 7581 individuals from seven countries in the Indian subcontinent. Using this dataset, this study investigates the inference of demographic attributes. Including nationality, religion, and gender, from personal names across the Indian subcontinent. The dataset encompasses seven different nationalities, five major religions, and two genders associated with each name. The dataset was constructed from Wikipedia biographies, public records, and social media profiles and was subsequently refined using cultural naming patterns to accurately label demographic attributes. Five models were used: three based on machine learning (LightGBM, XGBoost, and Support Vector Machine), one based on deep learning (Bidirectional Long Short‐Term Memory), and one based on transformers (XLM‐RoBERTa). The applied models yielded class‐wise accuracy, and the analysis indicates that the Support Vector Machine (SVM) achieved the highest accuracy across all three classes within the five models, with 83.23% for nationality, 92.94% for religion, and 92.67% for gender. In addition, the applied transformer‐based XLM‐RoBERTa model also showed excellent performance, particularly in religion and gender prediction. To assess robust accuracy, various performance metrics (precision, recall, and F 1 ‐score) were also computed for each class under the model. The analysis demonstrates that, for predicting demographic attributes using different AI paradigm models, character‐level textual features extracted from names are highly effective and that traditional machine learning models can outperform more complex deep learning architectures in name‐based demographic prediction.
Zitieren
APA
Sultana, S., Atib, Z. B. M., Redoy, M. O. F., Rahman, M. A., Uddin, M. S., & Habib, M. T. (2026). Comparative AI ‐Based Framework for Name‐Based Demographic Inference in the Indian Subcontinent. Engineering Reports, 8(6). https://doi.org/10.1002/eng2.70902
Vancouver
Sultana S, Atib ZBM, Redoy MOF, Rahman MA, Uddin MS, Habib MT. Comparative AI ‐Based Framework for Name‐Based Demographic Inference in the Indian Subcontinent. Engineering Reports. 2026;8(6). doi:10.1002/eng2.70902
BibTeX
@article{Sultana2026,
author = {Sultana, Sherin and Atib, Zaed Bin Monir and Redoy, Md. Omur Faruk and Rahman, Md. Ataur and Uddin, Mohammad Shorif and Habib, Md. Tarek},
title = {Comparative AI ‐Based Framework for Name‐Based Demographic Inference in the Indian Subcontinent},
journal = {Engineering Reports},
year = {2026},
volume = {8},
number = {6},
doi = {10.1002/eng2.70902}
}
EUVIMED