EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

PSO-Based Hybrid Lexical-Semantic Feature Selection and Ensemble Learning Framework for Urdu Hate Speech Detection

Haseeb Ullah, Husnain Saleem, Asia Kanwal, Muhammad Javed, Hamid Masood Khan, Kiran Hanif

BRAIN. Broad Research in Artificial Intelligence and Neuroscience · 16.09.2026 · 17 (3) · 43

DOI: 10.70594/brain/17.3/3 ↗

Kurz gefasst

Fachartikel in der Zeitschrift BRAIN. Broad Research in Artificial Intelligence and Neuroscience, online erschienen am 16.09.2026. Verfasst von Haseeb Ullah und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Automated Urdu hate-speech detection remains challenging because annotated resources are limited, orthography varies, and harmful meaning depends on both lexical and contextual cues. This study evaluates a leakage-safe lexical-semantic framework combining 5,000-dimensional word/bigram TF-IDF features with 768-dimensional multilingual sentence embeddings, followed by binary particle swarm optimisation (BPSO) and classical ensemble learning. Experiments used the Urdu text and binary labels from MMHS11K: 8,800 balanced training records and an untouched balanced test set of 2,200 records. BPSO selected 2,849 of 5,768 hybrid dimensions, reducing dimensionality by 50.61%. On the official test set, Stacking with the complete hybrid representation achieved Macro-F1 = 0.8468 and ROC-AUC = 0.9265; the PSO-selected representation achieved Macro-F1 = 0.8391 and ROC- AUC = 0.9196. After mask selection, classifier fitting time decreased by 52.24%, excluding sentence-embedding extraction and BPSO search. Leakage-safe nested five-fold validation showed a similar trade-off: fitting time decreased by 52.64%, with Macro-F1 = 0.8444 ± 0.0066 after PSO versus 0.8492 ± 0.0110 without PSO. Soft Voting and Stacking were statistically comparable on selected features. Three-seed transformer aggregates performed better: Urdu-RoBERTa achieved the highest Macro-F1 (0.8836), and XLM-R achieved the highest ROC-AUC (0.9529). Holm-corrected exact McNemar tests confirmed that transformer aggregates significantly outperformed PSO Stacking. Overall, BPSO approximately halves representation size and downstream classical-model fitting time, with a small predictive decrease that was not statistically significant relative to complete-hybrid Stacking after correction.

Zitieren

APA

Ullah, H., Saleem, H., Kanwal, A., Javed, M., Khan, H. M., & Hanif, K. (2026). PSO-Based Hybrid Lexical-Semantic Feature Selection and Ensemble Learning Framework for Urdu Hate Speech Detection. BRAIN. Broad Research in Artificial Intelligence and Neuroscience, 17(3), 43. https://doi.org/10.70594/brain/17.3/3

Vancouver

Ullah H, Saleem H, Kanwal A, Javed M, Khan HM, Hanif K. PSO-Based Hybrid Lexical-Semantic Feature Selection and Ensemble Learning Framework for Urdu Hate Speech Detection. BRAIN. Broad Research in Artificial Intelligence and Neuroscience. 2026;17(3):43. doi:10.70594/brain/17.3/3

BibTeX
@article{Ullah2026,
  author = {Ullah, Haseeb and Saleem, Husnain and Kanwal, Asia and Javed, Muhammad and Khan, Hamid Masood and Hanif, Kiran},
  title = {PSO-Based Hybrid Lexical-Semantic Feature Selection and Ensemble Learning Framework for Urdu Hate Speech Detection},
  journal = {BRAIN. Broad Research in Artificial Intelligence and Neuroscience},
  year = {2026},
  volume = {17},
  number = {3},
  pages = {43},
  doi = {10.70594/brain/17.3/3}
}
Kostenlos starten →