Artificial Intelligence-Based Healthcare Sector Disease Diagnosis Model

Abstract

Because of recent developments in IoT, cloud computing, and AI, the traditional healthcare system has been replaced by a smart healthcare system. Improvements to healthcare may be made via the use of cutting-edge technology like the Internet of Things and artificial intelligence. Multiple possibilities exist for the healthcare industry as a result of the merging of IoT and AI. This study contributes to the literature by introducing a novel model for the diagnosis of diseases in a smart healthcare system that is based on the convergence of artificial intelligence and the internet of things. Focusing on AI and IoT convergence methods, this paper aims to create a model for the diagnosis of cardiovascular and metabolic disorders. The provided model consists of many steps, including data collection, pre-processing, classification, and fine-tuning of parameters. Bearables and sensors are examples of IoT devices that facilitate data collecting, and artificial intelligence approaches use this information to aid in the diagnosis of sickness. The suggested technique diagnoses diseases using a Cascaded Long Short-Term Memory (CSO-CLSTM) model based on the Crow Search Optimization algorithm. We use CSO to fine-tune the CLSTM model's 'weights' and 'bias' parameters so that we can better classify medical data. Also, the isolating Forest (Forest) approach is used to filter out anomalous data in this study. The diagnostic results of the CLSTM model may be greatly enhanced by using CSO. Utilizing medical records, the CSO-LSTM model's efficacy was verified. Maximum accuracy rates of 96.16% and 97.26% were achieved in the experimental diagnosis of heart disease and diabetes, respectively, using the given CSO-LSTM model. As a result, the suggested CSO-LSTM model may be used in advanced healthcare systems for the purpose of illness diagnostics.