Research And Design Of A Medical Mega Data Analysis System Based On Hadoop

With the rapid development of computer application technology popularization and computer network,the process of informatization construction of medical industry all over the world also continues to accelerate,so the sub system of electronic medical record(EMR)and image archiving and communication subsystem(PACS)based hospital information system(HIS)has been an unprecedented wide range of applications.However the traditional single node data warehouse as the main storage mode of hospital information system(HIS)due to the limitation of hardware cost,for large amounts of unstructured data processing to meet the performance bottleneck,it is difficult to achieve two-way extended storage capacity and computing power.This paper draws on the latest research fruits in the field of current data processing,design a based on Hadoop medical data analysis system,in order to better meet the hospital demand for large amount of medical data integration process and quantitative analysis.Aiming at the actual operation of the internal medical data information construction in hospitals,first proposed the medical data analysis of Hadoop system structure based on;then according to the characteristics of Hadoop framework,combined with Hive,Hbase and Sqoop and three party tools,the cooperation between the system and HIS,realize all kinds of the medical data stored in medical databases of traditional single node to HDFS distributed storage management functions;statistics in diagnosis and massive Intelligence Aided raw data analysis in the process of research,a Map/Reduce algorithm with variable Key design,combined with the algorithm to establish the specific structure of the Hash table,the storage problem is solved on the Key-Value value different types of data;finally,this paper based on the encapsulation and code using Java language,then through the smoke Take 5 million copies of electronic medical records for data analysis,intelligent medical aided diagnostic template is generated,and data by recording experiments results show that the system than the traditional single node medical databases have higher data statistical analysis efficiency.The system with the implementation of low cost,safe and reliable data storage,scalability strong and statistical data analysis speed and many other advantages can support for doctors in the clinical diagnosis of decision-making can also be for hospitals to large medical group data mining depth and the system to raise the analysis to provide the technical support,in a certain extent improve the intelligent hospital,information integration construction level,promote the medical industry technology development.