基于ICD-10诊断编码的慢性病并发症聚类算法.PDF

TOPIC 专题 37 基于ICD-10诊断编码的 慢性病并发症聚类算法 1,2,3, 4 5 1,2 王晓霞 ，蒋伏松，王宇 ，熊贇 1. 复旦大学计算机科学技术学院，上海 201203；2. 上海市数据科学重点实验室，上海 201203； 3. 西北师范大学计算机科学与工程学院，甘肃 兰州 730070；4. 上海市第六人民医院，上海 200233； 5. 上海市普陀区疾病预防控制中心，上海 200333 摘要 慢性病与其相关并发症关系的研究，对患者以及医学研究都有重要意义。电子病历中记录的患者就诊数据 为研究目标慢性病与其并发症的关系提供了数据基础，其中面临的挑战之一在于既需要使用临床医生的领 域知识对并发症进行标注，又不希望给医生增加过多负担。设计了一种采用分组策略的基于ICD-10诊断编 码的慢性病并发症半监督聚类方法，以实现在较少的医生专家参与下对慢性病并发症归类。真实糖尿病患 者电子医疗记录数据集上的实验结果表明提出的算法是实用且有效的。 关键词 半监督学习；聚类算法；慢性病并发症；ICD-10诊断编码 doi: 10.11959/j.issn.2096-0271.2018028 Chronic disease complications clustering based on ICD-10 diagnoses code 1,2,3 4 5 1,2 WANG Xiaoxia , JIANG Fusong , WANG Yu , XIONG Yun 1. School of Computer Science, Fudan University, Shanghai 201203, China 2. Shanghai Key Laboratory of Data Science, Shanghai 201203, China 3. College of Computer Science and Engineering, Northwest Normal University, Lanzhou 730070, China 4. Shanghai Sixth People’s Hospital, Shanghai 200233, China 5. Shanghai Putuo District Center for Disease Control and Prevention, Shanghai 200333, China Abstract Study on the relationship between the chronic disease and the corresponding complications has great theoretical significance and applicable value for patients and clinical medicine. In order to utilize healthcare electronic record more reasonably, preprocessing was needed according to prior medical knowledge for chronic disease complication. The challenge of this work is that medical knowledge should be exploited to label the corresponding complications. To meet these challenges and assist physicians in labeling complications of a target chronic disease, a semi-supervised chronic disease compl