Deep Learning-Driven Novel Strategy For Iot-Cloud-Based Smart Healthcare Monitoring System

Abstract

Internet of Things (IoT)-enabled mobile healthcare apps provide millions of people with web-based amenities and health assistance. Massive quantities of big data are handled by such applications, which also make utilization of cloud computing (CC) for open and secure preservation. In the rapidly evolving field of healthcare used for better well-being and way of life, CC is vital. This work suggested IoT-cloud-driven healthcare system for tracking and recognizing critical illnesses to provide consumers with better services than digital medical applications. This paper introduces a novel quality-aware feature-tuned deep belief network (QF-DBN)for diabetes illness diagnosis and severity assessment. The generation of disease knowledge from healthcare information presents challenges for conventional DBN. QF-DBN combines health-related variables, develops effective Restricted Boltzmann Machine (RBMs), and trains RBMs to discover deep disease-relevant characteristics. In the first stage, medical data is created by utilizing medical sensors and open-source datasets to identify individuals who are likely to have severe diabetes. For enhancing the data by removing noisy data, min-max normalization is used in the second stage of data preprocessing. The third stage involves determining the set of features through the use of linear discriminant analysis (LDA). The proposed QF-DBN framework predicts diabetic illness in the fourth stage. In the fifth stage, the effectiveness of the suggested model is examined; along with a comparison between the suggested and current models. The research utilized SPSS and t-test for statistical analysis. The results of the experiment show that in a smart health surveillance system, the QF-DBN strategy performed better than other approaches.