TY - GEN
T1 - UA-CRNN
T2 - 28th ACM International Conference on Information and Knowledge Management, CIKM 2019
AU - Tan, Qingxiong
AU - Ma, Andy Jinhua
AU - Ye, Mang
AU - Yang, Baoyao
AU - Deng, Huiqi
AU - Wong, Vincent Wai Sun
AU - Tse, Yee Kit
AU - Yip, Terry Cheuk Fung
AU - Wong, Grace Lai Hung
AU - Ching, Jessica Yuet Ling
AU - Chan, Francis Ka Leung
AU - Yuen, Pong Chi
N1 - Publisher Copyright:
© 2019 Association of Computing Machinery.
PY - 2019/11/3
Y1 - 2019/11/3
N2 - Accurate prediction of mortality risk is important for evaluating early treatments, detecting high-risk patients and improving healthcare outcomes. Predicting mortality risk from the irregular clinical time series data is challenging due to the varying time intervals in the consecutive records. Existing methods usually solve this issue by generating regular time series data from the original irregular data without considering the uncertainty in the generated data, caused by varying time intervals. In this paper, we propose a novel Uncertainty-Aware Convolutional Recurrent Neural Network (UA-CRNN), which incorporates the uncertainty information in the generated data to improve the mortality risk prediction performance. To handle the complex clinical time series data with sub-series of different frequencies, we propose to incorporate the uncertainty information into the sub-series level rather than the whole time series data. Specifically, we design a novel hierarchical uncertainty-aware decomposition layer (UADL) to adaptively decompose time series into different sub-series and assign them proper weights according to their reliabilities. Experimental results on two real-world clinical datasets demonstrate that the proposed UA-CRNN method significantly outperforms state-of-the-art methods in both short-term and long-term mortality risk predictions.
AB - Accurate prediction of mortality risk is important for evaluating early treatments, detecting high-risk patients and improving healthcare outcomes. Predicting mortality risk from the irregular clinical time series data is challenging due to the varying time intervals in the consecutive records. Existing methods usually solve this issue by generating regular time series data from the original irregular data without considering the uncertainty in the generated data, caused by varying time intervals. In this paper, we propose a novel Uncertainty-Aware Convolutional Recurrent Neural Network (UA-CRNN), which incorporates the uncertainty information in the generated data to improve the mortality risk prediction performance. To handle the complex clinical time series data with sub-series of different frequencies, we propose to incorporate the uncertainty information into the sub-series level rather than the whole time series data. Specifically, we design a novel hierarchical uncertainty-aware decomposition layer (UADL) to adaptively decompose time series into different sub-series and assign them proper weights according to their reliabilities. Experimental results on two real-world clinical datasets demonstrate that the proposed UA-CRNN method significantly outperforms state-of-the-art methods in both short-term and long-term mortality risk predictions.
KW - Convolutional Recurrent Neural Network
KW - Machine Learning
KW - Mortality Risk Prediction
KW - Time Series Decomposition
KW - Uncertainty-Aware Prediction
UR - http://www.scopus.com/inward/record.url?scp=85075441682&partnerID=8YFLogxK
U2 - 10.1145/3357384.3357884
DO - 10.1145/3357384.3357884
M3 - Conference proceeding
AN - SCOPUS:85075441682
T3 - International Conference on Information and Knowledge Management, Proceedings
SP - 109
EP - 118
BT - CIKM 2019 - Proceedings of the 28th ACM International Conference on Information and Knowledge Management
PB - Association for Computing Machinery (ACM)
Y2 - 3 November 2019 through 7 November 2019
ER -