“…Many studies have been carried out to tackle the challenging problem of automatic MeSH indexing based on different principles, such as k -nearest neighbor (KNN) ( Trieschnigg et al , 2009 ), Naive Bayes ( Jimeno-Yepes et al , 2012b ), support vector machine (SVM) ( Jimeno-Yepes et al , 2012a ), Learning to Rank (LTR) ( Huang et al , 2011a ; Liu et al , 2015 ; Mao and Lu, 2013 ), deep learning ( Jimeno-Yepes et al , 2014 ; Rios and Kavuluru, 2015 ) and multi-label learning ( Liu et al , 2015 ; Tsoumakas et al , 2013 ). MeSHLabeler is a state-of-the-art automatic MeSH indexing algorithm, which won the first place in the large-scale MeSH indexing task of both BioASQ2 and BioASQ3 competition ( http://bioasq.org ) ( Liu et al , 2015 ; Tsatsaronis et al , 2015 ).…”