临床实验室检验知识的数据挖掘和表达  

The data mining and knowledge presentation in clinical laboratory test domain

在线阅读下载全文

作  者:张玉海[1] 苏海霞[2] 孙丽君[1] 尚磊[1] 王锐[1] 杨显君[1] 

机构地区:[1]第四军医大学卫生统计学教研室,陕西省西安市710032 [2]第四军医大学流行病学教研室

出  处:《中国医院统计》2012年第4期241-244,共4页Chinese Journal of Hospital Statistics

基  金:基金项目:国家自然科学基金(30901241)陕西省科技攻关项目(2009K18-01)

摘  要:目的探索基于自组织映射(SOM)的数据挖掘方法从历史积累的临床实验室检验数据中发现知识的能力,并尝试利用本体方法将该领域相关知识进行标准化表达。方法本研究的基本思想是采用自组织映射神经网络技术,从过去积累的数据中挖掘出专家经验基础上的临床检验项目的应用规律(知识)。以2009--2011年西安市两所综合医院的内科门诊患者就诊资料为训练样本,采用自组织映射方法建立门诊患者的聚类模型。采用通用的本体构建原则和方法,提取领域内相关概念及相互关系,建立临床实验室检验领域本体框架。结果SOM网络将就诊患者聚为8个类别,每个类别都具有较明显的临床意义,即每个类别中的患者具有相似的临床实验室检验项目的应用规律。聚类结果可以涵盖69.73%的患者。患者性别、年龄、3年累计临床检验项目数、疾病特征对聚类模型的贡献较大。构建了临床实验室检验领域本体框架,该框架的构建过程符合本体的建立准则,本体中的概念及其关系相对完整。结论自组织映射方法对门诊患者聚类效果较好,说明采用数据挖掘方法从历史数据中发现临床实验室检验项目的应用规律是可行的。采用本体方法将该领域知识进行标准化表达是一种有益的尝试。Objective To explore the ability of a data mining method based on self-organizing maps (SOM) to discover knowledge in historical data, and to test the feasibility to present the domain knowledge using ontology method. Methods The basic idea of this model is that mining the application rules of clinical laboratory test items (knowledge) from the past accumula- ted data on the basis of expert experience using self-organizing map neural network technology. The training sample is outpatients' data from two hospitals in Xi' an in 2009-2011. Clustering models of patients were built using the SOM method. The general prin- ciple and method of constructing ontology were used to extract relative concepts of the domain and construct the ontology frame- work. Results Self-organizing maps network clustered the outpatients into eight categories. Each category has obvious clinical meanings, namely the outpatients in each category have similar regular patterns in applying clinical laboratory test items. The out- put of clustering covered 69.73% outpatients. The sex, age, the numbers of clinical laboratory test in three years and disease characteristics of patients made great contribution to the clustering models. An ontology framework of clinical laboratory test do- main was constructed. The constructing process was accord with general principle and the concepts and their correlations was in- tegrated in a certain extent. Conclusion The satisfied SOM clustering output for outpatients indicated that using data mining method to discover the application rules of clinical test items from the past accumulated data is feasible. Using ontology method to present domain knowledge is an instructive attempt.

关 键 词:临床实验室检验 自组织映射 数据挖掘 本体知识表达 

分 类 号:R446[医药卫生—诊断学]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象