A general feature-weighting function for classification problems
作者：
Highlights：
• The salience of a feature can be changed given different queries.
• Multi-modal problems need a flexible weighting function to deal with modality.
• Proposing a differentiable, flexible, and general weighting function.
• The proposed is optimized for k-NN algorithm and assessed through experiments.
• The experiment results confirm the effectiveness of proposed weighting function.
摘要
•The salience of a feature can be changed given different queries.•Multi-modal problems need a flexible weighting function to deal with modality.•Proposing a differentiable, flexible, and general weighting function.•The proposed is optimized for k-NN algorithm and assessed through experiments.•The experiment results confirm the effectiveness of proposed weighting function.
论文关键词：Machine learning,Dynamic feature weighting,Weighting function,Nearest neighbor,Multi-modal weighting
论文评审过程：Received 20 July 2016, Revised 8 December 2016, Accepted 9 December 2016, Available online 9 December 2016, Version of Record 22 December 2016.
论文官网地址：https://doi.org/10.1016/j.eswa.2016.12.016
