Fall risk assessment through a synergistic multi-source DNN learning model
作者：
Highlights：
• We introduce the Synergy block, which successfully learns complementarities among multiple data sources.
• Our Synergy block identifies dependencies between fall conditions and environments/physical attributes of monitored adults.
• We build a multi-source LSTM model with the Synergy block, which takes advantage of synergy among multiple data forms.
• We use the learned synergy to identify possible novel fall risk factors in older population.
摘要
•We introduce the Synergy block, which successfully learns complementarities among multiple data sources.•Our Synergy block identifies dependencies between fall conditions and environments/physical attributes of monitored adults.•We build a multi-source LSTM model with the Synergy block, which takes advantage of synergy among multiple data forms.•We use the learned synergy to identify possible novel fall risk factors in older population.
论文关键词：Deep learning,Fall risk factors,Attention,Multi-source learning,Multi-label classification
论文评审过程：Received 17 August 2021, Revised 7 March 2022, Accepted 8 March 2022, Available online 18 March 2022, Version of Record 18 March 2022.
论文官网地址：https://doi.org/10.1016/j.artmed.2022.102280
