An elderly health monitoring system based on biological and behavioral indicators in internet of things

• Published in: Journal of ambient intelligence and humanized computing Vol. 14; no. 5; pp. 5085 - 5095
• Main Authors: Hosseinzadeh, Mehdi, Koohpayehzadeh, Jalil, Ghafour, Marwan Yassin, Ahmed, Aram Mahmood, Asghari, Parvaneh, Souri, Alireza, Pourasghari, Hamid, Rezapour, Aziz
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.05.2023; Springer Nature B.V
• Subjects: Artificial Intelligence; Biomonitoring; Classifiers; Computational Intelligence; Data analysis; Decision making; Decision trees; Engineering; Internet of Things; Machine learning; Multilayer perceptrons; Multilayers; Older people; Original Research; Quality of service; Recall; Robotics and Automation; Sensors; Telemedicine; User Interfaces and Human Computer Interaction; Internet of things; Health monitoring system; Data mining; Smart elderly care
• DOI: 10.1007/s12652-020-02579-7
• ISSN: 1868-5137

Advancement of sensor technologies has conducted to the rapid evolution of platforms, tools and approaches such as Internet of Things (IoT) for developing behavioral and physiological monitoring systems. Nowadays, According to growing number of elderlies living alone without their relatives scattered over the wide geographical areas, it is significantly essential to track their health function status continuously. In this paper, an IoT-based health monitoring system is proposed to check vital signs and detect biological and behavioral changes via smart elderly care technologies. It provides a health monitoring system for the involved medical teams to continuously monitor and assess a disabled or elderly’s behavioral activity as well as the biological parameters, applying sensor technology through the IoT devices. In this approach, vital data is collected via IoT monitoring objects and then, data analysis is carried out through different machine learning methods such as Decision Tree (J48), Sequential Minimal Optimization (SMO), Multi-Layer Perceptron (MLP) and Naïve Bayes (NB) classifiers for detecting the level of probable risks of elderly’s physiological and behavioral changes. The experimental results confirm that the SMO, MLP and NB classifiers meet approximately close performance considering the accuracy, precision, recall, and f-score factors. However, the J48 method shows the highest performance for health function status predicting in our scenario with 99%, of accuracy and precision, 100% of recall and 97% of f-score. Moreover, the J48 performs with the lowest execution time in comparison to the other applied classifiers.