Multi-label learning deals with objects associated with multiple class labels, and aims to induce a predictive model which can assign a set of relevant class labels for an unseen instance. Since each class might possess its own characteristics, the strategy of extracting label-specific features has been widely employed to improve the discrimination process in multi-label learning, where the predictive model is induced based on tailored features specific to each class label instead of the identical instance representations. As a representative approach, LIFT generates label-specific features by conducting clustering analysis. However, its performance may be degraded due to the inherent instability of the single clustering algorithm. To improve this, a novel multi-label learning approach named SENCE (stable label-Specific features gENeration for multi-label learning via mixture-based Clustering Ensemble) is proposed, which stabilizes the generation process of label-specific features via clustering ensemble techniques. Specifically, more stable clustering results are obtained by firstly augmenting the original instance repre-sentation with cluster assignments from base clusters and then fitting a mixture model via the expectation-maximization (EM) algorithm. Extensive experiments on eighteen benchmark data sets show that SENCE performs better than LIFT and other well-established multi-label learning algorithms.

Yi-Bo Wang;Jun-Yi Hang;Min-Ling Zhang

School of Computer Science and Engineering,Southeast University,Nanjing 210096,China

自动化学报（英文版）

[1]Yi-Bo Wang;Jun-Yi Hang;Min-Ling Zhang-.Stable Label-Specific Features Generation for Multi-Label Learning via Mixture-Based Clustering Ensemble)[J].自动化学报（英文版）,2022(07):1248-1261

SENCE,gENeration

Stable,Label,Features,Generation,Multi,Learning,via,Mixture,Based,Clustering,Ensemble,learning,deals,objects,associated,multiple,class,labels,aims,predictive,model,which,can,relevant,unseen,instance,Since,each,might,possess,its,own,characteristics,strategy,extracting,specific,features,has,been,widely,employed,improve,discrimination,process,where,induced,tailored,instead,identical,representations,representative,approach,LIFT,generates,by,conducting,clustering,analysis,However,performance,may,degraded,due,inherent,instability,single,To,this,novel,named,stable,mixture,proposed,stabilizes,generation,ensemble,techniques,Specifically,more,results,are,obtained,firstly,augmenting,original,assignments,from,clusters,then,fitting,expectation,maximization,EM,Extensive,experiments,eighteen,benchmark,data,sets,show,that,performs,better,than,other,well,established,algorithms

0.550153