IoMT Assisted Monitoring and Voice-Based Food Recommendation System Using Deep Learning Model

Abstract

Many individuals in the contemporary world are afflicted by a wide range of illnesses and diseases. Providing a diet recommendation on short notice is notoriously challenging. An AI-powered, cloud-based medical automation system has the potential to extend human life, prevent the spread of illness, and improve people's general level of health. Using real-time data from biomedical sensors on 50 patients, this study presents a deep learning approach to construct an IoMT-aided health information system capable of autonomously detecting which food should be supplied to individuals. Provided information is sent to the cloud, and the user will get updates on their health status from the cloud. With the user's current health situation in mind, the system formulates a voice-based dietary suggestion for maximum effectiveness. The suggested technique is improved upon by using deep learning algorithms such as recurrent neural networks (RNNs), multilayer perceptrons (MLPs), and Long Short-Term Memories (LSTMs). It was determined by comparing the precision, recall, accuracy, and F1-measures of a number of different deep learning methods that the LSTM methodology is the most effective. Using an LSTM deep learning model, we were able to get an accuracy of 89.9%. For the permitted class, we get an F1-measure of 0.86 s, recall of 0.87, and accuracy of 0.89.