Survey on multi-label learning
Zhixi Li · Jisuanji yingyong yanjiu · 2014
Multi-label learning framework is proposed for ambiguity object which is ubiquitous in real-word.This paper provided a comprehensive survey on research situation and progress of multi-label learning.Firstly,it presented definition and solving strategies of multi-label learning.Secondly,it divided the state-of-the-art multi-label learning techniques into two categories,i.e.problem transformation and algorithm adaptation.In addition,it emphatically described the learning principles of each category.Thirdly,it discussed definitions and functions of various evaluation measures.Finally,it summarized several valuable research directions under the background of multi-label learning.