Online disease risk monitoring using DEWMA control chart
作者：
Highlights：
• A robust DEWMA chart is suggested to monitor disease risk factors sequentially.
• Different kernel functions are used to estimate the risk factors of stroke patients.
• DEWMA perform better and gives less signals in detecting risk than the EWMA chart.
• Models like mixed effect, time series, and non-Gaussian are used to generate covariates.
• Epanechinkov and cosine kernels perform better with small smoothing parameters.
• Tricube and biweight kernels perform better with large smoothing parameters.
摘要
•A robust DEWMA chart is suggested to monitor disease risk factors sequentially.•Different kernel functions are used to estimate the risk factors of stroke patients.•DEWMA perform better and gives less signals in detecting risk than the EWMA chart.•Models like mixed effect, time series, and non-Gaussian are used to generate covariates.•Epanechinkov and cosine kernels perform better with small smoothing parameters.•Tricube and biweight kernels perform better with large smoothing parameters.
论文关键词：Average time to signal,DEWMA chart,Online disease risk monitoring,Kernels,Stroke patients
论文评审过程：Received 13 January 2021, Revised 6 April 2021, Accepted 15 April 2021, Available online 22 April 2021, Version of Record 8 May 2021.
论文官网地址：https://doi.org/10.1016/j.eswa.2021.115059
