Labelling strategies for hierarchical multi-label classification techniques

作者：

Highlights：

• We select single or multiple thresholds for hierarchical multi-label classifiers.

• Selecting multiple thresholds often yield better label sets in lesser time.

• We show that optimising H-loss tends to favor empty label sets.

• Multiple threshold selection is preferred for micro F-measure and HMC-loss.

• Imitating training set properties is a competitive approach to optimise HMC-loss.

摘要

Highlights•We select single or multiple thresholds for hierarchical multi-label classifiers.•Selecting multiple thresholds often yield better label sets in lesser time.•We show that optimising H-loss tends to favor empty label sets.•Multiple threshold selection is preferred for micro F-measure and HMC-loss.•Imitating training set properties is a competitive approach to optimise HMC-loss.

论文关键词：Hierarchical multi-label classification,Threshold optimisation,Hierarchical loss,HMC-loss,F-measure

论文评审过程：Received 16 October 2015, Revised 29 January 2016, Accepted 24 February 2016, Available online 4 March 2016, Version of Record 12 April 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2016.02.017