MEDOOP: Medical Health Information System Based on Hadoop

Abstract

Medical information exchange is an imminent stage of medical informatics development. Integrating the administration feature, the structure of HIE platform accepts a centralized architecture which has to serve a big volume of medical information and abut multiple types of query and information extraction for centralized design. Cloud computing is a technological requirement. In this paper, we put forward a medical information platform based on Hadoop, which is named after Medoop. The current Medoop make effective use of HDFS to store the integrated CDA records more effectively, arrange the content data in CDA record as per the continuous business queries and calculate the statistic distributedly in Map Reduce model. The aim of Medoop is to implement a complete platform for medical data to be stored, exchanging and using the element in Hadoop environment. The project presents Medical Hadoop which consists of cloud server and predicts wether the patient is affected by diabetes or not. For predicting this we use Naïve Bayes algorithm. Naïve Bayes is provided with static dataset in its training phase and trained accordingly for detection. Naïve Bayes is provided with normalized dataset using k-means(or any other normalization algorithm).