医疗数据的清洗以及知识挖掘  

Data cleaning of medical data and knowledge mining

作　　者：吴辰文 郑恒[1] 张烨[1] 朱建东[1] 

机构地区：[1]兰州交通大学电子与信息工程学院,甘肃兰州730070

出　　处：《安徽大学学报（自然科学版）》2016年第1期23-29,共7页Journal of Anhui University(Natural Science Edition)

基　　金：兰州市科技计划基金资助项目(2009-1-5);甘肃省自然科学基金资助项目(1308RJZA111)

摘　　要：数据挖掘或数据分析在生物医学中与其他调查领域不同,因为在生物医学领域这些数据很复杂,他们资源不同,且每一个医师对同一个诊断记录都有他们自己的解释.分析医疗数据的特征,研究数据的清洗,目的是为了挖掘有价值的知识.实验显示,所提出的方法比朴素贝叶斯网络模式更有效.Data mining or data analysis in biomedicine is different from other research fields,because the data in biomedical are heterogeneous,and they are from different sources.Moreover,each physician might have his own interpretation with the same clinical records.In this paper,we analyzed the features of medical data,and studied data cleaning for medical data in order to mine valuable knowledge.Experiments showed that the proposed method was more efficient than the baseline Bayesian network model.

关 键 词：数据清理 医疗数据 知识挖掘 贝叶斯网络 

分 类 号：U416[交通运输工程—道路与铁道工程]