Fatemah Husain

Dr. Fatemah Husain

orcid

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
19 January 2022 in ACM Transactions on Asian and Low-Resource Language Information Processing
Verified
Journal Article
Open access
FHFatemah Husain
Fatemah Husain
OUOzlem Uzuner
Ozlem Uzuner
Preprocessing of input text can play a key role in text classification by reducing dimensionality and removing unnecessary content. This study aims to investigate the impact of preprocessing on Arabic offensive language classification. We explore six preprocessing techniques: conversion of emojis to Arabic textual labels, normalization of different forms of Arabic letters, normalization of selected nouns from dialectal Arabic to Modern Standard Arabic, conversion of selected hyponyms to hypernyms, hashtag segmentation, and basic cleaning such as removing numbers, kashidas, diacritics, and HTML tags. We also experiment with raw text and a combination of all six preprocessing techniques. We apply different types of classifiers in our experiments including traditional machine learning, ensemble machine learning, Artificial Neural Networks, and Bidirectional Encoder Representations from Transformers (BERT)-based models to analyze the impact of preprocessing. Our results demonstrate significant variations in the effects of preprocessing on each classifier type and on each dataset. Classifiers that are based on BERT do not benefit from preprocessing, while traditional machine learning classifiers do. However, these results can benefit from validation on larger datasets that cover broader domains and dialects.

8 February 2024 in IEEE Access
Verified
Journal Article
Open access
FHFatemah Husain
Fatemah Husain
HAHana Alostad
Hana Alostad
HOHalima Omar
Halima Omar
The available dialectal Arabic linguistic resources are very limited in their coverage of Arabic dialects, particularly the Kuwaiti dialect. This shortage in linguistic resources creates struggles for researchers in the Natural Language Processing (NLP) field and limits the development of advanced linguistic analytical and processing tools for the Kuwaiti dialect. Many other low-resource Arabic dialects are still not explored in research due to the challenges faced during the annotators’ recruitment process for dataset labeling. This paper proposes a weak supervised classification system to solve the problem of recruiting human annotators called "q8SentiLabeler". In addition, we developed a large dataset consisting of over 16.6k posts serving sentiment analysis for the Kuwaiti dialect, covering several themes and time-frames to remove any bias that might affect its content. Furthermore, we evaluate our dataset using multiple traditional machine-learning classifiers and advanced deep-learning language models to test its performance. Results demonstrate the positive potential of "q8SentiLabeler" to replace human annotators with a 93% for the pairwise percent agreement and a 0.87 for Cohen’s kappa coefficient. Using the ARBERT model on our dataset, we were able to achieve 89% accuracy in the system’s performance.

17 June 2025 in IEEE Access
Verified
Journal Article
Open access
ZAZainab AlMeraj
Zainab AlMeraj
FHFatemah Husain
Fatemah Husain
RARosa I. Arriaga
Rosa I. Arriaga
This paper examines cross-lingual and cross-cultural variations in disability discourse on social media, specifically on X (formerly Twitter). Despite the growing presence of disability-related discussions online, empirical evidence remains limited regarding the perception, accessibility, and advocacy for people with disabilities in non-English-speaking countries of the Global South. Utilizing Natural Language Processing (NLP), this research analyzes three years of Arabic and English disability-related posts, incorporating two hashtags, #peoplewithdisabilities and #disability. The study investigates posting behaviors, engagement patterns, sentiment, and empathy-driven linguistic differences, while also assessing the cultural factors shaping these narratives. The findings show that there are notable linguistic and thematic differences between the posts in English and Arabic across different continents. They also highlight how these relate to disability theories and offer insights for policy makers and advocates on how to effectively study, track, and design for disability awareness in the global south.

19 December 2023 in ACM Transactions on Asian and Low-Resource Language Information Processing
Verified
Journal Article
Open access
FHFatemah Husain
Fatemah Husain
Multiple attempts to develop systems for detecting online Arabic offensive language have been explored in previous studies. However, most of these attempts do not consider the variation of Arabic dialects, cultures, and offensive phrases. In contrast, this study aims to extract knowledge from multiple offensive language datasets to build a cross-dialect and culture knowledge-based repository. This knowledge-based repository is utilized to develop novel system architecture based on customizing the AraBERT model in a unique method to preserve dialectal knowledge and offensive cultural knowledge within the contextual word embedding of BERT architecture. Performance evaluation procedures consist of statistical evaluation metrics and a behavioral checklist. Results report more effective predictions by the customized model than the uncustomized one, particularly for offensive text. The customization process allows the model to gain more knowledge of informal text in general, and a better understanding of dialectal Arabic.

8 February 2026 in Social Network Analysis and Mining
Verified
Journal Article
Open access
PDF file
FHFatemah Husain
Fatemah Husain
This study addresses the critical gap in detecting offensive language within Kuwaiti Arabic (KA), a dialect extensively used in informal digital communication yet poorly supported by existing Natural Language Processing (NLP) systems. Given the increasing prevalence of abusive online content that threatens cybersecurity, safety, and social cohesion, the development of dialect-sensitive detection models is essential. Building upon prior work, this research constructs the first Kuwaiti offensive language data set by annotating an existing corpus, enabling the development of a specialized text classification system tailored to KA. The proposed system not only aims to accurately identify offensive content but also integrates sentiment analysis to explore the relationship between emotional expression and offensive language. Additionally, a detailed empathetic lexicon-based analysis examines linguistic patterns and emotional triggers that drive offensive communication in KA online discourse. The work includes comprehensive model evaluation, misclassification analysis, and deployment of a web-based demo. We also carried out an error analysis using both manual inspection and explainable AI approaches such as SHapley Additive exPlanations (SHAP). The findings underscore the importance of dialect-specific NLP resources for safeguarding digital spaces and enhancing cybersecurity strategies, while also contributing to a deeper understanding of the linguistic and emotional dynamics that underpin offensive language in Kuwaiti social media interactions.

5 July 2023
Verified
Conference Paper
Open access
ZAZainab AlMeraj
Zainab AlMeraj
FHFatemah Husain
Fatemah Husain
Existing research targeting disability on social media has looked primarily at the English language across the Global North. This has caused significant data gaps in the understanding of disability awareness and cultural mindsets elsewhere. This work presents a cross-language analysis of English and Arabic #disability tweets over a period of time. A dictionary of words was adopted to help in understanding the linguistics surrounding the term ”empathy”. The finding suggests that disability mentions in English tweets have slightly more negative empathetic emotions than those found in Arabic tweets. However, more investigations are needed to explore cultural variations towards empathy and disability, and how they differ linguistically in online conversations.