Artificial Intelligence Based Explainable Smart Healthcare: Integration and Analysis of IoT Data

Abstract

In recent years, great advances in technology have led to the popularisation of the Internet of Things (IoT) concept. Especially in the field of smart health, IoT has become one of the most preferred areas for researchers and practitioners thanks to the continuous monitoring of patient's health. The use of IoT sensor networks in healthcare plays an important role in services such as patient monitoring, chronic disease control and emergency response. This study aims to analyse the data obtained from IoT devices with machine learning (ML) algorithms using acceleration data of Parkinson's patients. Parkinson's disease is a chronic disease that greatly reduces the quality of life by affecting the nervous system and impairing motor skills. Analysing the movement states of Parkinson's patients is very important in terms of better understanding the disease and developing treatment methods. In this study, Decision Trees (DT), Lightweight Gradient Boosting Machine (LGBM) and Random Forest (RF) algorithms were used to predict freezing states in the data obtained with acceleration sensors of Parkinson's patients and LGBM showed the highest performance with %98 accuracy. LIME explainable artificial intelligence algorithm was used to determine the factors that most affect the model performance. This study demonstrates the potential benefits of integrating IoT and ML technologies to improve the quality of life of Parkinson's patients and increase the effectiveness of healthcare services. © The Institution of Engineering & Technology 2024.