Joint learning for voice based disease detection

作者：

Highlights：

• A model based on losses coupling is proposed for voice based disease detection.

• Regression targets are relaxed to obtain more discriminant representation.

• Low-rank constraints are imposed to make the representation more compact.

• Key factors for the methods are separately analyzed by experiment.

摘要

•A model based on losses coupling is proposed for voice based disease detection.•Regression targets are relaxed to obtain more discriminant representation.•Low-rank constraints are imposed to make the representation more compact.•Key factors for the methods are separately analyzed by experiment.

论文关键词：Joint learning,Ridge regression,Low-rank regression,ϵ-dragging technique,Voice based pathology detection

论文评审过程：Received 7 March 2018, Revised 1 September 2018, Accepted 27 September 2018, Available online 11 October 2018, Version of Record 23 November 2018.

论文官网地址：https://doi.org/10.1016/j.patcog.2018.09.013