A hybrid cost-sensitive ensemble for imbalanced breast thermogram classification
作者：
Highlights：
• Novel cost-sensitive hybrid ensemble for efficient breast thermogram classification.
• Evolutionary algorithm for simultaneous feature selection and classifier weighting.
• A combined criterion for the ensemble training algorithm.
• Estimating cost parameter with the usage of ROC analysis.
• Embedded pruning procedure for discarding irrelevant classifiers.
摘要
Highlights•Novel cost-sensitive hybrid ensemble for efficient breast thermogram classification.•Evolutionary algorithm for simultaneous feature selection and classifier weighting.•A combined criterion for the ensemble training algorithm.•Estimating cost parameter with the usage of ROC analysis.•Embedded pruning procedure for discarding irrelevant classifiers.
论文关键词：Classifier ensemble,Imbalanced classification,Cost-sensitive classification,Ensemble pruning,Evolutionary algorithm,Breast cancer detection,Thermogram
论文评审过程：Received 11 May 2013, Revised 15 July 2015, Accepted 23 July 2015, Available online 31 July 2015, Version of Record 14 November 2015.
论文官网地址：https://doi.org/10.1016/j.artmed.2015.07.005
