Latent ranking analysis using pairwise comparisons in crowdsourcing platforms
作者：
Highlights：
• We study how to learn multiple latent rankings from pairwise comparisons.
• We propose a novel probabilistic model to capture the process of pairwise comparison.
• We develop an efficient inference algorithm to learn multiple latent rankings.
• We also investigate active learning problem considering crowdsourcing platforms.
• Experiments with synthetic and real life data show the effectiveness of our algorithms.
摘要
Highlights•We study how to learn multiple latent rankings from pairwise comparisons.•We propose a novel probabilistic model to capture the process of pairwise comparison.•We develop an efficient inference algorithm to learn multiple latent rankings.•We also investigate active learning problem considering crowdsourcing platforms.•Experiments with synthetic and real life data show the effectiveness of our algorithms.
论文关键词：Learning to rank,Pairwise comparison,Active learning,Crowdsourcing
论文评审过程：Received 8 April 2015, Revised 3 October 2016, Accepted 15 October 2016, Available online 20 October 2016, Version of Record 14 November 2016.
论文官网地址：https://doi.org/10.1016/j.is.2016.10.002
