LLM-Based Enhanced Clustering for Low-Resource Language: An Empirical Study

Talha Farooq Khan¹, Majid Hussain¹, Muhammad Arslan², Muhammad Saeed¹, Lal Khan^3,*, Hsien-Tsung Chang^4,5,6,*
1 Department of Computer Science, The University of Faisalabad, Faisalabad, 38000, Pakistan
2 Department of Computer Science, The University of Southern Punjab, Multan, 60000, Pakistan
3 Department of AI and SW, Gachon University, Seongnam, 13120, Republic of Korea
4 Department of Artificial Intelligence, Chang Gung University, Linkou, Taoyuan, 333, Taiwan
5 Department of Computer Science and Information Engineering, Chang Gung University, Linkou, Taoyuan, 333, Taiwan
6 Center for Artificial Intelligence in Medicine, Chang Gung Memorial Hospital at Linkou, Linkou, Taoyuan, 333, Taiwan
* Corresponding Authors: Lal Khan. Email: lalkhan@gachon.ac.kr; Hsien-Tsung Chang. Email: smallpig@cgu.edu.tw
(This article belongs to the Special Issue: Applied NLP with Large Language Models: AI Applications Across Domains)

Computer Modeling in Engineering & Sciences https://doi.org/10.32604/cmes.2025.073021

Received 09 September 2025; Accepted 24 November 2025; Published online 08 December 2025

Download PDF

Abstract

Text clustering is an important task because of its vital role in NLP-related tasks. However, existing research on clustering is mainly based on the English language, with limited work on low-resource languages, such as Urdu. Low-resource language text clustering has many drawbacks in the form of limited annotated collections and strong linguistic diversity. The primary aim of this paper is twofold: (1) By introducing a clustering dataset named UNC2025 comprises 100k Urdu news documents, and (2) a detailed empirical standard of Large Language Model (LLM) improved clustering methods for Urdu text. We explicitly evaluate the behavior of the 11 multilingual and Urdu-specific embeddings on 3 different clustering algorithms. We carefully evaluated our performance based on a set of internal and external measurements of validity. We discover the best configuration of the mBERT embedding with the HDBSCAN algorithm that attains a new state-of-the-art performance with a high score of external validity of 0.95. This new LLM method has created a new strong standard of Urdu text clustering. Importantly, the results confirm the strength and high scalability of the LLM-generated embeddings towards the ability to generalise the fine, subtle semantics needed to discover topics in low-resource settings and open the door to novel NLP applications in underrepresented languages.