Active learning with multi-criteria decision making systems
作者：
Highlights：
• We apply multi-criteria decision making systems to the field of active learning.
• The distance between preference preorders is used as the metric measure.
• Unlabeled examples are evaluated by the dominated and dominating indices.
• We implement the learning framework under multiple-instance environment.
• Experimental comparisons show superior performance of the proposed work.
摘要
Highlights•We apply multi-criteria decision making systems to the field of active learning.•The distance between preference preorders is used as the metric measure.•Unlabeled examples are evaluated by the dominated and dominating indices.•We implement the learning framework under multiple-instance environment.•Experimental comparisons show superior performance of the proposed work.
论文关键词：Active learning,Multiple-instance learning,Multi-criteria decision making,Support vector machine
论文评审过程：Received 19 October 2012, Revised 26 September 2013, Accepted 18 March 2014, Available online 27 March 2014.
论文官网地址：https://doi.org/10.1016/j.patcog.2014.03.011
