High Accurate Stroke Prediction Smart Necklace Device Based on IoT and Machine Learning Algorithm

• Published in: 2023 1st International Conference on Health Science and Technology (ICHST) pp. 1 - 6
• Main Authors: Phuc, Pham Hong, Toan, Vu Duc, Ha, Tran Quang, Nghia, Vu Huu, Minh Thu, Le Thi, Son, Bui Cao, Phuc, Truong Duc
• Format: Conference Proceeding
• Language: English
• Published: IEEE 28.12.2023
• Subjects: Accuracy; Fall Detection; IoT; Machine learning; Machine learning algorithms; Medical services; Predictive models; Radio frequency; Smart Necklace; Stroke Prediction; Support vector machines
• DOI: 10.1109/ICHST59286.2023.10565374

This article deals with the design and development of a smart necklace device that can warn users of a stroke problem. The device has been developed in accordance with loT standards, and it consists of a necklace paired with a mobile phone application that allows caretakers or medical doctors to readily identify the wearer's position and health index data. The device is able to collect an user's movement data such as acceleration, angular velocity, and electrocardiogram through sensors. The collected data is proceeded by using machine learning to predict whether a stroke occurs or not based on fall detection. Whenerver the wearer exhibits indicators of a sudden fall, the device will sound an alarm to alert individuals nearby for help, togother with sending notifying SMS messages to caregivers and medical doctors. The location of the wearer is conveniently tracked on maps accurately by using GPS sensor for emergency activities. Through experiments, prediction accuracy of over 94% was demonstrated for the smart necklace wearable device developed in this study.