{
    "Name": "MedQA-MA",
    "Dialect Subsets": [],
    "HF Link": "",
    "Link": "https://data.mendeley.com/datasets/v6gs7nsy9z/1",
    "License": "CC BY-NC 4.0",
    "Year": 2025,
    "Language": "ar",
    "Dialect": "Morocco",
    "Source": [
        "web pages",
        "books",
        "social media"
    ],
    "Domain": [
        "health"
    ],
    "Form": "text",
    "Annotation Style": [
        "machine annotation",
        "human annotation"
    ],
    "Description": "This dataset constitutes the first large-scale collection of medical question\u2013answer pairs in Moroccan Arabic,",
    "Volume": 108943.0,
    "Unit": "sentences",
    "Ethical Risks": "Low",
    "Provider": [
        "Sidi Mohamed Ben Abdellah University"
    ],
    "Derived From": [],
    "Paper Title": "MedQA-MA: A Moroccan Arabic medical question-answering dataset for virtual healthcare assistants and large language models",
    "Paper Link": "https://doi.org/10.1016/j.dib.2026.112537",
    "Script": "Arab",
    "Tokenized": false,
    "Host": "Mendeley Data",
    "Access": "Free",
    "Cost": "",
    "Has Splits": false,
    "Partial": false,
    "Tasks": [
        "question answering"
    ],
    "Venue Title": "Data in Brief",
    "Venue Type": "journal",
    "Venue Name": "Data in Brief",
    "Authors": [
        "Soufiyan Ouali",
        "Said El Garouani"
    ],
    "Affiliations": [
        "Sidi Mohamed Ben Abdellah University"
    ],
    "Abstract": "The healthcare domain constitutes a fundamental pillar of\nnational development, as maintaining population health not\nonly enhances citizens\u2019 quality of life but also generates sub-\nstantial economic benefits through increased productivity, in-\nnovation, and workforce participation. However, the health-\ncare industry faces numerous challenges and barriers that\nimpede universal access to medical services. In low- and\nmiddle-income countries, significant portions of the popu-\nlation forego medical consultations due to various socioe-\nconomic constraints, including prohibitive consultation fees,\nscheduling difficulties, and extended waiting periods. Conse-\nquently, there is an urgent need for innovative approaches\nto optimize healthcare delivery processes. Recent advances\nin artificial intelligence have demonstrated promising poten-\ntial in developing intelligent systems that address health-\ncare accessibility gaps. These innovations include medical\nchatbots, appointment booking systems, disease-prediction\nmodels, and psychiatric virtual assistants. However, such\ntechnological enhancements have predominantly focused on\nhigh-resource languages, while research in low-resource lan-\nguages, particularly Arabic, remains in its preliminary stages.\nThis disparity is especially pronounced in Arabic dialects,\nwhich differ substantially from Modern Standard Arabic in\nterms of vocabulary, syntax, and semantic structures. To\naddress this critical gap, we present the first comprehen-\nsive dataset for the Moroccan Arabic dialect in the health-\ncare domain. The MedQA-MA dataset comprises 108,943\nquestion-answer pairs in text format, with each pair cat-\negorized according to medical specialty. Including 23 dis-\ntinct medical specialties, this dataset serves multiple ap-\nplications, including sentiment analysis, specialty classifica-\ntion, question-answering systems, and the development of\nhuman-like medical chatbots. The dataset has been meticu-\nlously curated, annotated, and validated by qualified medical\nprofessionals, ensuring its reliability and clinical relevance for\ndeveloping realistic healthcare systems grounded in authen-\ntic medical interactions.\nThe MedQA-MA dataset is publicly available and freely ac-\ncessible at https://data.mendeley.com/datasets/v6gs7nsy9z/1,\nrepresenting a significant contribution to Arabic Natural Lan-\nguage Processing research in healthcare applications and fa-\ncilitating the development of culturally and linguistically ap-\npropriate medical AI systems for Arabic-speaking populations",
    "Added By": "Zaid Alyafeai"
}