Design and implementation of cyber-physical data collection system of risk factors mitigation and life expectancy prediction
Year: 2026
Type: conference-paper
Abstract: Life expectancy is a critical indicator of public health, influenced by multiple factors including behavioral and lifestyle choices such as physical activity, diet, smoking, and alcohol consumption. Existing health monitoring systems are often fragmented, and data collected by smart devices are difficult to integrate into analytical platforms. To address this, the proposed cyber-physical system (CPS) connects smart scales via Bluetooth to a custom Android application, automatically storing biometric data in a database for machine learning–based analysis. Machine learning models — Random Forest, Gradient Boosting, and Ridge Regression — were evaluated for life expectancy prediction using RMSE, MAE, and R2metrics. Gradient Boosting achieved the highest accuracy, with MAE=0.26 years, RMSE=0.80 years, and R2=1.00, while Random Forest and Ridge Regression also showed strong predictive performance. Feature importance analysis indicates that age and BMI are the strongest predictors of life expectancy. Additional lifestyle and biometric factors, such as physical activity and body composition, also contribute significantly. These findings highlight the value of analyzing multiple lifestyle and biometric factors to support personalized health assessments and guide interventions for improving life expectancy.
Authors: Asina Bopushova, Ruslan Isaev
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Keywords: life expectancy prediction, cyber-physical systems, smart scales, biometric data collection, health monitoring, lifestyle risk factors, body mass index, physical activity, body composition, gradient boosting, random forest, ridge regression, feature importance, Android application, Bluetooth data acquisition
Open Access status: hybrid
OpenAlex ID: W7212298272
