干眼症AI训练数据集

本数据集源自爱尔眼科自研的【爱尔病历中心】2022年8月1日至2025年6月30日的病历数据，数据范围涵盖个7省份，10家医院。经专家进行规范的病历数据标准设计，并对病历数据的准确性和完整性进行筛选，然后经过严格脱敏脱密、清洗、融合与结构化处理后，形成用于干眼症AI训练的高质量数据集，数据规模：6201份。该数据集以“AI干眼医生”核心模型训练为导向，系统性覆盖诊疗全链条，并整合病史、检查结果等多维度患者信息，内置分型依据、治疗方案等输出内容。可服务于干眼症临床诊断关键变量的识别、模型性能提升。基于此数据集训练的“AI干眼医生”大模型产品可显著减少诊疗主观差异，提升决策质量。随着数据与算法迭代，本数据集将持续驱动诊疗系统升级，推进干眼症诊疗标准化、智能化进程，并为真实世界研究提供关键支撑。

This dataset is derived from the medical record data from August 1, 2022 to June 30, 2025 of the self-developed Aier Medical Record Center of Aier Eye Hospital. The data covers 7 provinces and 10 hospitals. After the standardized design of medical record data standards by experts, screening of the accuracy and integrity of medical record data, and then undergoing strict de-identification, cleaning, integration and structuring processing, a high-quality dataset for dry eye syndrome AI training is formed, with a total of 6201 cases. Oriented towards the training of the core "AI Dry Eye Doctor" model, this dataset systematically covers the entire diagnosis and treatment workflow, integrates multi-dimensional patient information including medical history and examination results, and has built-in output contents such as classification criteria and treatment plans. It can be used for the identification of key variables in clinical diagnosis of dry eye syndrome and the improvement of model performance. The "AI Dry Eye Doctor" large language model (LLM) product trained on this dataset can significantly reduce subjective differences in diagnosis and treatment and improve decision-making quality. With the iteration of data and algorithms, this dataset will continue to drive the upgrading of diagnosis and treatment systems, promote the standardization and intelligentization of dry eye syndrome diagnosis and treatment, and provide critical support for real-world research.

搜集汇总

背景与挑战

背景概述

该数据集是一个专注于干眼症的AI训练数据集，包含6201条结构化病历数据，涵盖发病年龄、起病月份等关键字段，用于构建疾病知识图谱和训练AI模型，以支持多模态辅助诊疗系统的开发，实现干眼症的个性化治疗方案生成和临床决策辅助。