A hybrid machine learning approach to cerebral stroke prediction based on imbalanced medical dataset
作者：
Highlights：
• A hybrid machine learning approach is used to predict stroke via incomplete and imbalanced medical dataset.
• Instance selection of automated hyperparameter optimization(AutoHPO) successfully extracts hard-classified samples to reduce the imbalance ratio.
• Online-learning strategy is used to reweight each batch of the training set and optimize the parameters of the model based on validation loss rather than training loss as usual, which achieves impressive performance on imbalanced dataset.
摘要
•A hybrid machine learning approach is used to predict stroke via incomplete and imbalanced medical dataset.•Instance selection of automated hyperparameter optimization(AutoHPO) successfully extracts hard-classified samples to reduce the imbalance ratio.•Online-learning strategy is used to reweight each batch of the training set and optimize the parameters of the model based on validation loss rather than training loss as usual, which achieves impressive performance on imbalanced dataset.
论文关键词：Stroke prediction,Clinical decision,Class imbalance,Hybrid machine learning,AutoHPO
论文评审过程：Received 28 March 2019, Revised 12 August 2019, Accepted 6 September 2019, Available online 23 October 2019, Version of Record 11 November 2019.
论文官网地址：https://doi.org/10.1016/j.artmed.2019.101723
