The Medical Data Collection Mechanism Based On The Block Data

With the advent of the big data era,the data in medical and health field contains more and more value,and the demanding of data sharing and interoperability is getting more and more urgent.But these data are “island” and “strip” data,which are heterogeneous,difficult to flow and share.That has seriously affected the data to play its due value.In order to solve such problem,the theory of "block data" is proposed,which is the sum of data formed in a physical space or administrative area.We apply that idea to the medical information construction,taking patient diagnosis and treatment information as the object,and research the mechanism of data collection based on “block data” to break the isolation between data and realize data interconnection.The goal is to achieve cross-platform,cross-institutional medical data integration collection.By studying and analyzing the heterogeneous characteristics of diagnosis and treatment data,we found that semantic heterogeneity is the key issue.In response to this problem,ontology technology is used to solve the heterogeneity,so that diagnosis and treatment information can understand each other;According to the needs of collection and referring to the hybrid ontology semantic model construction method,the overall idea and model are designed to eliminate the semantic heterogeneity and achieve data collection independent of interface;According to the characteristics and the types of the data,the way of establishing the local semantic model of patient diagnosis and treatment data is analyzed and designed.We extract semantic model from relation database and XML file to establish the local ontology.It solves the heterogeneity due to form and language.By comparing the way of constructing domain ontology,the way to establish a global shared ontology is designed;By studying the mapping method between ontology,we adopt the method based on computing ontology concepts similarity,and design the overall process and model of establishme nt the mapping.We study and analyze the key problem of mapping,the similarity calculation method.An automatic weighting method is used to mix the multi-angle similarity calculation results to obtain more comprehensive and accurate results.The experiment result shows that it can receive better results and improve the accuracy.We implement the core tasks and select three actual data sources about patient diagnosis and treatment information to conduct the data collection model construction experiment.The experiment shows that ontology technology can effectively solve the semantic problem in medical data.The medical data collection method proposed can meet the current heterogeneous medical data collection requirements.