A Study on the Development of the Classification Models Between Fallers and Non-Fallers with Wearable Sensors and Questionnaires in Older Adults  
Author Jungyeon Choi

 
Co-Author(s) Seung Woo Lee; Kwang Yoon Song

 
Abstract Falls are a significant health concern among older adults, leading to adverse outcomes such as injury, loss of independence, and increased mortality. This study aims to develop and evaluate classification models to differentiate fallers from non-fallers using data collected from wearable sensors and self-reported questionnaires. Wearable sensors provide continuous monitoring of gait, balance, and activity levels, while questionnaires offer insights into fall history, health status, and psychological factors. A cohort of older adults was recruited, and their data were analyzed using machine learning algorithms to build predictive models. Various feature selection techniques were applied to enhance model performance, and cross-validation was employed to ensure robustness. The results demonstrated that combining sensor data with questionnaire responses significantly improves the accuracy of fall risk classification. These findings highlight the potential of integrating wearable technology and subjective assessments for early identification of fall risks in older adults, facilitating timely intervention and reducing fall-related incidents.

 
Keywords Risk of fall, Machine learning, Older adults, Gait analysis, Wearable sensor
   
    Article #:  DSBFI25-6
 
Proceedings of 3rd ISSAT International Conference on Data Science in Business, Finance and Industry
January 6-8, 2025 - Da Nang, Vietnam