A novel three-tier Internet of Things architecture with machine learning algorithm for early detection of heart diseases

Abstract

Among the applications enabled by the Internet of Things (IoT), continuous health monitoring system is a particularly important one. Wearable sensor devices used in IoT health monitoring system have been generating an enormous amount of data on a continuous basis. The data generation speed of IoT sensor devices is very high. Hence, the volume of data generated from the IoT-based health monitoring system is also very high. In order to overcome this issue, this paper proposes a scalable three-tier architecture to store and process such huge volume of wearable sensor data. Tier-1 focuses on collection of data from IoT wearable sensor devices. Tier-2 uses Apache HBase for storing the large volume of wearable IoT sensor data in cloud computing. In addition, Tier-3 uses Apache Mahout for developing the logistic regression-based prediction model for heart diseases. Finally, ROC analysis is performed to identify the most significant clinical parameters to get heart disease.