| With the continuous development of database technology and information processing technology,hospital information systems have not only greatly improved medical efficiency,but also accumulated a large amount of medical data with research value.However,due to the characteristics of phased construction and inconsistent construction standards in the development of hospital information systems,the processing and analysis of medical data has brought tremendous challenges.Therefore,by establishing a proprietary data warehouse and ETL(Extract-Transform-Load)tool in the medical field,multi-source heterogeneous data aggregation can be completed,thus providing data analysis decision support for medical institutions and promoting the development of hospital information systems.ETL is a collective term for data extraction,cleaning,conversion,and loading processes.Traditional ETL tools are oriented to the general domain.It is difficult for hospital data to deal with heterogeneous data of multiple data sources efficiently.This paper deeply analyzes the data characteristics of multiple heterogeneous information systems in hospitals,and the complex complex relationships between data cleaning,transformation,and loading.This leads to the need to resolve the aggregate issues of multi-source heterogeneous data in the follow-up data warehouse ETL process.According to the characteristics of data,this paper deeply studies the solution to guide the whole process of ETL through metadata,and proposes an ontology-based metadata management model.Through the idea of layering,the complex relationship between data is combed to solve multi-source heterogeneous data.The integration problem is to efficiently and accurately complete the follow-up data warehouse ETL process.At the same time,in the process of data transformation,the semi-structured data of the discharge summary is identified by named entity,the key medical data is extracted intelligently and intelligently,the dimension of patient data is increased,and complete data covering the patient’s full medical history is constructed.This thesis analyzes the requirements of the data warehouse for hospital follow-up services,completes the design and implementation of the follow-up data warehouse model and metadata management model,designs and implements metadata-based data warehouse ETL tools.At the same time,the practical application of the follow-up center in the hospital verified the availability and efficiency of the solution in the follow-up data warehouse ETL process. |
