huzaib-khan-23/augmented-clinical-notes

增强临床笔记数据集是一个包含30,000个三元组的扩展数据集，这些三元组来自不同的来源：真实的临床笔记（来自PMC-Patients数据集）、合成的患者-医生对话（由GPT 3.5生成）以及结构化的患者信息（由GPT-4生成）。该数据集用于训练MediNote-7B和MediNote-13B等临床笔记生成器。数据集的语言为英语，包含的字段有idx（唯一标识符）、note（可能被截断的临床笔记）、full_note（完整的临床笔记）、conversation（患者-医生对话）和summary（患者信息摘要，JSON格式）。数据集的主要用途是训练模型从临床笔记中提取结构化的患者信息，也可用于医疗保健领域的其他应用。

The Augmented Clinical Notes dataset is an extension of existing datasets containing 30,000 triplets from different sources: real clinical notes (from PMC-Patients dataset), synthetic patient-doctor conversations (generated by GPT 3.5), and structured patient information (generated by GPT-4). This dataset was used to train clinical note generators such as MediNote-7B and MediNote-13B. The dataset is in English and includes fields such as idx (unique identifier), note (possibly truncated clinical note), full_note (full clinical note), conversation (patient-doctor dialogue), and summary (patient information summary in JSON format). The primary use of the dataset is to train models to extract structured patient information from clinical notes, and it can also be used for other applications in the healthcare domain.