Hezam/AHD

AHD: Arabic healthcare dataset

Dataset Description

Title:"AHD: Arabic healthcare dataset"

Paper: AHD: Arabic healthcare dataset

Journal: Data in Brief

Publication year: October 2024

Repository: Mendeley Data

Languages

• The text in the dataset is in English.
• The text in the dataset is in Arabic.

Specifications

• Subject: Computer Science, Data Science, Health and medical sciences.
• Specific subject area: Arabic Language, Machine Learning, Health Informatics, Natural Language Processing, Text Generation, Text classification.
• Type of data: Text/String
• Data Format: Raw
• Data collection: The dataset was collected from the Altibbi website using web-scraping tools. Python, with the Requests and BeautifulSoup packages, was used to collect the data.
• Size: Consists of over 808k rows 90 variety categories.

Dataset Summary

With the soaring demand for healthcare systems, chatbots are gaining tremendous popularity and research attention. Numerous language-centric research on healthcare is conducted day by day. Despite significant advances in Arabic Natural Language Processing (NLP), challenges remain in natural language classification and generation due to the lack of suitable datasets. The primary shortcoming of these models is the lack of suitable Arabic datasets for training. To address this, authors introduce a large Arabic Healthcare Dataset (AHD) of textual data. The dataset consists of over 808k questions and answers across 90 categories, offered to the research community for Arabic computational linguistics. Authors anticipate that this rich dataset would make a great aid for a variety of NLP tasks on Arabic textual data, especially for text classification and generation purposes. Authors present the data in raw form. AHD is composed of main dataset scraped from medical website, which is Altibbi website. AHD is made public and freely available at http://data.mendeley.com/datasets/mgj29ndgrk/5.

Value of the Data

• AHD is the largest, to our knowledge, available and representative Arabic Healthcare Dataset (AHD) for a wide variety category.
• AHD offers up to ninety distinct categories, making it robust for accurate text categorization.
• AHD offers over 808k distinct questions and answers, making it robust for accurate healthcare systems and chatbots.
• In contrast with the few small available datasets, AHD's size makes it a suitable corpus for implementing both classical as well as deep learning models.

Supported Tasks and Leaderboards

The dataset supports the task of text generation, Text classification, Chatbot, Question answering and Word embedding.

Dataset Structure

Source Data

The data was sourced from online counseling and therapy platform. The raw data can be found altibbi.

Data Files

• 'AHD.xlsx:' file contains dataset in excel format, which includes the question, answer, and category in Arabic.
• 'AHD_english.xlsx:' file contains dataset in excel format, which includes the question, answer, and category translated to English.
• 'Distribution of Question and Answer per category.xlsex:' shows the distribution of the data set by category.

Data Instances

A data instance includes a 'Question', 'Answer' and 'Category'. 'Question' contains the question asked by a use , 'Answer' contains the corresponding answer provided by a docto , and 'Category' lable for question and answer from 90 variety categories.

Data Fields

• 'Question': a string containing the question asked by a user
• 'Answer': a string containing the corresponding answer provided
• 'Category': a lable for questions from 90 variety categories.

Data Splits

• The dataset has no predefined splits.
• The dataset adopted the annotation of each question as it appeared on its source website, Altibbi.
• Users can create their own splits as needed.

Distribution of Question and Answer

The AHD consists of more than 808k rows 90 variety categories

Ethics Statement

• 'Terms of Service (ToS):' Authors have considered and followed the source website Altibbi's ToS, privacy laws, and user consents.
• 'Privacy:' The authors have anonymized all participant data and confirm that all the data is non-sensitive.
• 'Scraping policies:' There are not specific scraping policies for the source website Altibbi.
• 'Copyright:' The authors have read and followed the ethical requirements for publication in Data in Brief and confirmed that the current work does not involve any type of human studies, animal studies, or data gathered using social media. All data belongs to the source website Altibbi through user consent, and it's open to the public, so ethical approval has not been sought. The data adopted the annotation of each question and answer as appeared on the source website Altibbi and the distribution of data per category. We can confirm that this manuscript adheres to ethical publishing standards.