Data-driven decision model based on dynamical classifier selection
作者：
Highlights：
• A decision model based on dynamical classifier selection (DCS) is proposed.
• DCS is implemented based on data similarity and predictive accuracy.
• Explainable decisions are generated from predictions of selected classifier.
• The proposed model is used to aid the diagnosis of thyroid nodules.
• The model is compared with traditional decision model and representative DCS methods.
摘要
•A decision model based on dynamical classifier selection (DCS) is proposed.•DCS is implemented based on data similarity and predictive accuracy.•Explainable decisions are generated from predictions of selected classifier.•The proposed model is used to aid the diagnosis of thyroid nodules.•The model is compared with traditional decision model and representative DCS methods.
论文关键词：Data-driven decision model,Dynamical classifier selection,Explainable decision-making,Interval-valued numbers,Diagnoses of thyroid nodules
论文评审过程：Received 17 February 2020, Revised 20 July 2020, Accepted 31 October 2020, Available online 13 November 2020, Version of Record 24 December 2020.
论文官网地址：https://doi.org/10.1016/j.knosys.2020.106590
