Improved Random Forest And Cuckoo Search Optimization Based Hybrid Approach For Healthcare Monitoring Systems

Abstract

Wearable sensors are one of the recent advances in the field of Remote Patient Monitoring (RPM), which are capable of recognizing the activities of patients, including their movements. However, the continuous monitoring of patients using wearable sensors tends to create a vast amount of medical data. Existing approaches based on monitoring the healthcare system are insufficient for extracting the required information from patients and lack accuracy. To overcome these issues related to patient healthcare monitoring, this study proposes an improved healthcare monitoring system using an Improved Random Forest (IRF). The IRF approach selects better parameters from every architectural model and provides a better pre-trained model to monitor the patient’s health conditions. The hyperparameters of the selected features were fine-tuned using the Improved Cuckoo Search (ICS) algorithm. The experimental outcomes indicate that the proposed method achieved a better accuracy of 99.45%, which is comparatively higher than the existing Federated Learning-based Person Movement Identification (FL-PMI) and FedStack with 97.78% and 98.11%, respectively.