🤖 AI Summary
This study addresses the limited accuracy of existing fall-risk prediction models for older adults by proposing a novel multimodal assessment framework that integrates motion data (accelerometry) with non-motor clinical features (e.g., age, comorbidities). Using multimodal data from 146 older adults, we systematically evaluated multiple machine learning models and identified Bayesian ridge regression as the top-performing method when leveraging fused data (MSE = 0.6746; R² = 0.9941), significantly outperforming unimodal baselines. Our key contributions are twofold: (1) We provide the first empirical validation that non-motor factors contribute independently and substantially to fall-risk prediction—beyond motion-derived metrics alone; and (2) we establish an interpretable, clinically actionable assessment paradigm grounded in “multisource data fusion + Bayesian modeling,” enabling precise, quantifiable risk estimation to support early clinical intervention.
📝 Abstract
This study investigates fall risk prediction in older adults using various machine learning models trained on accelerometric, non-accelerometric, and combined data from 146 participants. Models combining both data types achieved superior performance, with Bayesian Ridge Regression showing the highest accuracy (MSE = 0.6746, R2 = 0.9941). Non-accelerometric variables, such as age and comorbidities, proved critical for prediction. Results support the use of integrated data and Bayesian approaches to enhance fall risk assessment and inform prevention strategies.