Work

MarsadLab at AraHealthQA: Hybrid Contextual–Lexical Fusion with AraBERT for Question and Answer Categorization

Year: 2025

Type: conference-paper

Abstract: This paper presents the MarsadLab submission to Track 1 of the AraHealthQA 2025 Shared Task, addressing two subtasks: (A) multi-label question categorization and (B) multi-label answer categorization in Arabic mental health discourse.Our approach employs a hybrid contextual-lexical fusion architecture built on AraBERTv2, enriched with task-specific handcrafted features such as lexical indicators, linguistic cues, and domain-informed keyword signals.On the official test set, the system achieved a weighted F1 score of 0.55 (Jaccard 0.41) for Task A and 0.79 (Jaccard 0.67) for Task B.

Source: Arabic Natural Language Processing Workshop (WANLP)

DOI: 10.18653/v1/2025.arabicnlp-sharedtasks.33

Authors: Mabrouka Bessghaier, Shimaa Ibrahim, Md. Rafiul Biswas, Wajdi Zaghouani

Institution: Northwestern University in Qatar

Language: English

Topic: Topic Modeling

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

OpenAlex ID: W4416037820

FWCI:

Cites:

Cited by:

Related to: 10