Multi-label ensemble based on variable pairwise constraint projection

Abstract

Abstract:Multi-label classification has attracted an increasing amount of attention in recent years. To this end, many algorithms have been developed to classify multi-label data in an effective manner. However, they usually do not consider the pairwise relations indicated by sample labels, which actually play important roles in multi-label classification. Inspired by this, we naturally extend the traditional pairwise constraints to the multi-label scenario via a flexible thresholding scheme. Moreover, to improve the generalization ability of the classifier, we adopt a boosting-like strategy to construct a multi-label ensemble from a group of base classifiers. To achieve these goals, this paper presents a novel multi-label classification framework named Variable Pairwise Constraint projection for Multi-label Ensemble (VPCME). Specifically, we take advantage of the variable pairwise constraint projection to learn a lower-dimensional data representation, which preserves the correlations between samples and labels. Thereafter, the base classifiers are trained in the new data space. For the boosting-like strategy, we employ both the variable pairwise constraints and the bootstrap steps to diversify the base classifiers. Empirical studies have shown the superiority of the proposed method in comparison with other approaches.

Multi-label feature selection based on max-dependency and min-redundancy

“…To date, a number of multi-label learning algorithms have been developed [15,18,21,40,42]. Depending on whether executing learning algorithm from multi-label data directly or not, these methods can be classified into two categories, i.e., problem transformation and algorithm adaption.…”

Using tensor products to detect unconditional label dependence in multilabel classifications

“…In [9] and [24] the approach was an extension of multiclass classification. Other extensions have been done from different settings; this is the case of Bayesian learners [33], nearest neighbors [34], logistic regression [15], decision trees [29] and ensembles [14].…”

“…In all cases the input space was R p+1 with p ∈ {10, 25, 50, 100}, see (14). The size m of the set of labels varied in {100, 150, 200}.…”

Multi-view based multi-label propagation for image annotation

“…But most existing work in the line of multi-label propagation suffer (or partially suffer) from the disadvantage that they consider each label independently when handling the multi-label propagation problem [2,6].…”