Design And Implementation Of Hadoop Based Stroke Assistant Diagnosis Platform

With the developing of medical informatization, traditional medicine has gradually become modernizational,and the hospital information system(HIS) has also been more and more popular.Nowadays,most hospitals have their own disease database,these databases detail the clinical informations and treatments of stroke patients.The stroke medical systems currently used have 3 main problems:1.the stroke medical data increases sharply,so the system need to do expansion work frequently.Moreover,the data is easily to be lost during the transfer process.2. most hospitals use their own systems,it makes the stroke medical data scattering.However,the treatment of stroke always during a long process.When the patient change hospital midway,the doctors from other hospital can not obtain the previous information of patient,so it is difficult for them to take the most effective treatment.3.the system just manage data for low level operation such as recording,modifying,deleting and searching.It is unable to make use of these data and find out valuable information from them.Therefore, this article puts forward a kind of stroke assistant diagnosis platform based on Hadoop, it aims to solve the problems above.The main contents of the platform include:1.with the full use of HBase system architecture and data storage mode, we can store the medical data from different hospitals in a same distribute database which based on the distributed filesytem (HDFS) of hadoop.It not only make the data expansion and scattering no longer be a problem,but also ensuret the security of data.2.using the multi-level architecture of J2EE and design mode of MVC,which makes the business response,business processing and result views separately,the platform offer the medical staff good service.3.base on MapReduce framework of Hadoop,the platform improve the tranditonal serial Apriori algorithm to a parallel mode.With the optimized algorithm,it mines the mass stroke medical data.The minging result can provide scientific evidence for clinical treatment and research of stroke,and improve the level of diagnosis of this disease.