noone-m/shifaa-processed

Shifaa阿拉伯语医疗咨询数据集（分类处理版）是一个经过处理的阿拉伯语医疗咨询数据集，用于单标签分类任务。该数据集包含患者提出的医疗问题及其对应的主要医疗专业类别。数据经过清洗处理，包括从层次标签中提取主类别、移除有冲突类别的问题、去重以及分层拆分（训练集约81%，验证集9%，测试集10%）。数据集包含三个特征：Question（问题文本）、Main Category（主类别），并分为train、validation和test三个split。

Shifaa Arabic Medical Consultations - Processed for Classification is a processed dataset of Arabic medical consultations for single-label classification tasks. The dataset contains patients medical questions along with their corresponding main medical specialty categories. The data has been cleaned by extracting main categories from hierarchical labels, removing questions with conflicting categories, deduplicating questions, and performing stratified splitting (~81% train, 9% val, 10% test). The dataset includes two features: Question (text of the question) and Main Category (main category), and is divided into train, validation, and test splits.