BrownBoost Classifier-Based Bloom Hash Data Storage for Healthcare Big Data Analytics

Abstract

Healthcare big data analytics is the process of collecting and analyzing a large volume of patient data to find out useful information. Big data analytics has a number of challenges in many fields including cloud healthcare systems. The healthcare industry significantly creates large volumes of patient data. Most of the recent research works happen on big data analytics-enabled business models for increasing the prediction accuracy to reduce the risk level of patients. However, storage of data is a major concern, and data need to be accessed effectively across various locations in the distributed environment. Our aim is to develop a BrownBoost Classifier-Based Bloom Hash Data Storage (BBC-BHDS) mechanism for storing and accessing the healthcare data from various locations in distributed environment with minimum space usage and in less time. Initially, a large volume of data (i.e., patient data) are collected based on certain features (parameters), and then classified the input using BrownBoost Classification (BBC) algorithm. BrownBoost employs a non-convex potential loss function and uses the base SVM classifier for classifying the patient data. Experiment of the proposed BBC-BHDS mechanism is carried out on number of data files. The results show that the proposed BBC-BHDS mechanism is more efficient with respect to classification accuracy, false positive rate, space complexity, and data accessing time in comparison with existing methods.