This paper introduces the Nakba Lexicon, a comprehensive dataset derived from the poetry collection Asifa ‘Ala al-Iz‘aj (Sorry for the Disturbance) by Istiqlal Eid, a Palestinian poet from El-Birweh. Eid’s work poignantly reflects on themes of Palestinian identity, displacement, and resilience, serving as a resource for preserving linguistic and cultural heritage in the context of post-Nakba literature. The dataset is structured into ten thematic domains, including political terminology, memory and preservation, sensory and emotional lexicon, toponyms, nature, and external linguistic influences such as Hebrew, French, and English, thereby capturing the socio-political, emotional, and cultural dimensions of the Nakba. The Nakba Lexicon uniquely emphasises the contributions of women to Palestinian literary traditions, shedding light on often-overlooked narratives of resilience and cultural continuity. Advanced Natural Language Processing (NLP) techniques were employed to analyse the dataset, with fine-tuned pre-trained models such as ARABERT and MARBERT achieving F1-scores of 0.87 and 0.68 in language and lexical classification tasks, respectively, significantly outperforming traditional machine learning models. These results highlight the potential of domain-specific computational models to effectively analyse complex datasets, facilitating the preservation of marginalised voices. By bridging computational methods with cultural preservation, this study enhances the understanding of Palestinian linguistic heritage and contributes to broader efforts in documenting and analysing endangered narratives. The Nakba Lexicon paves the way for future interdisciplinary research, showcasing the role of NLP in addressing historical trauma, resilience, and cultural identity.

Among many potential subjects studied in Sentiment Analysis, widespread offensive and abusive language on social media has triggered interest in reducing its risks on users; children in particular. This paper centres on distinguishing between offensive and abusive language detec- tion within Arabic social media texts through the employment of various machine and deep learning techniques. The techniques include Naïve Bayes (NB), Support Vector Machine (SVM), fastText, keras, and RoBERTa XML multilingual embeddings, which have demon- strated superior performance compared to other statistical machine learning methods and dif- ferent kinds of embeddings like fastText. The methods were implemented on two separate corpora from YouTube comments totalling 47K comments. The results demonstrated that all models, except NB, reached an accuracy of 82%. It was also shown that word tri-grams en- hance classification performance, though other tuning techniques were applied such as TF-IDF and grid-search. The linguistic findings, aimed at distinguishing between offensive and abu- sive language, were consistent with machine learning (ML) performance, which effectively classified the two distinct classes of sentiment: offensive and abusive.