A Study on Semantic Data Acquisition and Large-scale Data Processing in Electronic Health Record Environment

Abstract

With developments of a medical practice environment and a computer technology, EMR (Electronic Medical Record) system computerized human healthcare records and documents into easily exchangeable formats. The improved medical information environment around medical fields accompanies methods to represent those medical information into formally structured data with interoperability, so that, the data can be exploited at various applications and services. Development of medical terminology system guaranteeing interoperability of diverse medical records and further research structuring medical data through the terminology systems have been preceded recently. For structuring the medical data and keeping them to be with interoperable characteristics, recently information technology suggests a novel data representation method named Semantic Web. The semantic web technology secures an interoperable use of different data from diverse information system by representing data with references to domain ontology which is a set of conceptualization of domain knowledge. To exploit the semantic web technology at the EHR (Electronic Healthcare Record) system, two main research issues are followed

developments of a method for acquisition of semantic data at the medical treatment field, and a method for processing the obtained semantic data with large scale. This thesis paper investigates on the two research issues for realization of EHR environment along with the semantically described and interoperable data. This paper describes a CNL (Controlled Natural Language)-based guided look-ahead editor might help users select proper words that meet his intended but vague notions without proper knowledge on the sentence structures. The CNL editor also permits grammar definitions through CFG-LD that includes both sequential and semantic views on sentence structures, so that, the editor are employed for diverse medical document with different description standards. Therefore, this paper provides a data processing method for the huge size of semantically structured data obtained from the proposed CNL editor through scalable parallelism like Map-Reduce framework. With a cloud computing environment providing Map-Reduce cluster, the proposed method provides the efficient data pre-processing method for SPARQL queries used at the application services for analyzing and querying on the semantically structured data.