Personalized prediction of drug efficacy for diabetes treatment via patient-level sequential modeling with neural networks
作者：
Highlights：
• Sequential modeling approach is applied to prediction modeling for type 2 diabetes.
• Sequential dependences in multiple records of patients are incorporated into modeling.
• Previous records of each patient are directly utilized to predict prescription efficacy.
• Prediction accuracy is improved for those patients with multiple previous records.
• Application can assist clinical decision support for patients with type 2 diabetes.
摘要
•Sequential modeling approach is applied to prediction modeling for type 2 diabetes.•Sequential dependences in multiple records of patients are incorporated into modeling.•Previous records of each patient are directly utilized to predict prescription efficacy.•Prediction accuracy is improved for those patients with multiple previous records.•Application can assist clinical decision support for patients with type 2 diabetes.
论文关键词：Type 2 diabetes mellitus,Drug failure prediction,Sequential modeling,Personalized prediction,Neural network
论文评审过程：Received 25 September 2017, Revised 15 January 2018, Accepted 15 February 2018, Available online 23 February 2018, Version of Record 16 March 2018.
论文官网地址：https://doi.org/10.1016/j.artmed.2018.02.004
