中医大模型辩证

数据集介绍
评测数据基于医院脱敏病历构建，共1500条数据。数据分为训练集、验证集和测试集，数据量分别为800、200和500。本任务仅公开训练集数据和无标签的验证集数据，测试集数据不公开。

数据由json格式给出，数据集包含以下内容：

TCM-TBOSD-train.json: 训练集标注数据。
TCM-TBOSD-test-A.json: A榜测试集（验证集）。
TCM-TBOSD-A.json: A榜提交示例。
TCM-TBOSD-test-B.json: B榜测试集（测试集）。B榜测试集不公开。
数据集申请
1.在阿里云天池完成报名，数据集可在天池直接下载获取。

标注数据的字段信息说明
ID：患者入院的唯一id
性别：男或女
职业：患者的职业信息，如职员、退（离）休人员等
年龄：患者的年龄。
婚姻：描述婚姻状况，如已婚、未婚等
病史陈述者：入院时描述患者身体状况的人员与患者本人的关系，如患者本人
发病节气：患者出现病情时所处于的节气，如清明、小雪等
主诉：患者在就诊时向医生描述的最主要、最直接的不适或症状，用一句简短的文本概括描述，通常是患者就医的主要原因
症状：患者入院时所表现出的主要症状和体征的概述
中医望闻切诊：医师对患者进行“望”、“闻”、“切”后，对患者状态的描述
病史：包括现病史、既往史、个人史、婚育史、家族史

Dataset Introduction
The evaluation dataset is constructed based on de-identified hospital medical records, totaling 1500 entries. The data is divided into training set, validation set, and test set, with 800, 200, and 500 samples respectively. Only the training set and unlabeled validation set data are publicly available for this task, while the test set data is not disclosed.

The data is provided in JSON format, and the dataset includes the following files:
TCM-TBOSD-train.json: Annotated training set data.
TCM-TBOSD-test-A.json: Test Set A (Validation Set).
TCM-TBOSD-A.json: Submission example for Test Set A.
TCM-TBOSD-test-B.json: Test Set B (Official Test Set). Test Set B is not publicly disclosed.

Dataset Application
1. Complete registration on Alibaba Cloud Tianchi, and the dataset can be downloaded directly from the Tianchi platform.

Annotation Field Instructions
ID: Unique identifier for the patient's hospital admission
Gender: Male or Female
Occupation: Professional information of the patient, such as staff, retired personnel, etc.
Age: The patient's age.
Marital Status: Describes the patient's marital status, such as married, unmarried, etc.
History Reporter: The relationship between the person describing the patient's medical condition at admission and the patient themselves, e.g., the patient themselves
Onset Solar Term: The solar term when the patient developed the illness, such as Qingming, Xiaoxue (Minor Snow), etc.
Chief Complaint: The most prominent and direct discomfort or symptom that the patient describes to the doctor during consultation, summarized in a short text, usually the main reason for the patient's visit
Symptoms: Overview of the main symptoms and signs exhibited by the patient upon admission
Traditional Chinese Medicine (TCM) Inspection, Auscultation & Olfaction, and Palpation: Description of the patient's condition made by the physician after performing TCM inspection, auscultation & olfaction, and palpation
Medical History: Includes present illness history, past medical history, personal history, marital and reproductive history, and family history

搜集汇总

背景与挑战

背景概述

该数据集是一个用于中医大模型辩证任务的数据集，基于医院脱敏病历构建，共包含1500条数据，分为训练集、验证集和测试集。数据提供了详细的患者信息字段，如主诉、症状和中医望闻切诊等，旨在支持中医辩证的自动化和模型训练。