DARL: EFFECTUAL DEEP ADAPTIVE REINFORCEMENT LEARNING MODEL ENABLED SECURITY AND ENERGY-EFFICIENT HEALTHCARE SYSTEM IN INTERNET OF THINGS WITH THE AID OF MODIFIED MANTA RAY FORAGING OPTIMIZATION APPROACH

Smart healthcare systems are evolving due to the Internet of Things (IoT) and sensor-body area network (SBAN) technology. These advancements enable the collection, storage, and evaluation of individual patient data, enabling timely and effective treatment provision. However, the challenge lies in physicians analyzing vast patient information to provide accurate diagnoses and treatments. To address these issues, more robust and smart medical systems are needed. Reinforcement learning (RL) has shown its ability to deliver sophisticated information and reliability, aiming to enhance moral decision-making through constructive criticism and communication. Therefore, a methodology of deep RL for a smart healthcare monitoring system (HMS) in an IoT is provided in this paper. Initially, the sensor data are taken from various sources of data. Then, the collected data are forwarded to the medical decision-making system using a deep adaptive reinforcement learning (DARL)-based disease prediction model. The DARL-based disease prediction model uses the sensor data to analyze and predict potential health issues in patients. By continuously learning and adapting based on new data, the system can provide more accurate and timely diagnoses. This innovative approach to healthcare monitoring shows great promise in improving patient outcomes and revolutionizing the way healthcare is delivered. Here, the parameters are optimized using the modified Manta-ray foraging optimization (MMRFO) to enhance the effectiveness. The MMRFO algorithm helps to fine-tune the DARL-based disease prediction model by finding the best combination of parameters to improve its performance. It has the potential to transform the healthcare industry by enabling earlier detection of diseases and more personalized treatment plans for patients. The combination of DARL and MMRFO showcases the power of artificial intelligence in revolutionizing healthcare delivery and ultimately improving patient outcomes. Finally, the developed DARL model offered the disease-predicted outcome. The performance of the suggested model is validated with various conventional systems to prove its effectiveness. The accuracy of the proposed MMRFO-DARL is 95.31%, and the comparison models like GSM, LWAMCNet, RNN, and RL are given the accuracy to be 79.82%, 83.33%, 87.37%, and 89.06%, respectively. The result proved that the suggested model is sufficient to sense and process information of the patients via the medical decision-supporting system. By utilizing this model to analyze data and provide personalized treatment plans, healthcare professionals can offer more targeted care to their patients. The validation of the model’s performance against traditional systems further solidifies its effectiveness and potential to revolutionize healthcare delivery.