Abstract:Purpose Image classification is becoming a supporting technology in several image-processing tasks. Due to rich semantic information contained in the images, it is very popular for an image to have several labels or tags. This paper aims to develop a novel multi-label classification approach with superior performance. Design/methodology/approach Many multi-label classification problems share two main characteristics: label correlations and label imbalance. However, most of current methods are devoted to either model label relationship or to only deal with unbalanced problem with traditional single-label methods. In this paper, multi-label classification problem is regarded as an unbalanced multi-task learning problem. Multi-task least-squares support vector machine (MTLS-SVM) is generalized for this problem, renamed as multi-label LS-SVM ((MLS)-S-2-SVM). Findings Experimental results on the emotions, scene, yeast and bibtex data sets indicate that the (MLS)-S-2-SVM is competitive with respect to the state-of-the-art methods in terms of Hamming loss and instance-based F1 score. The values of resulting parameters largely influence the performance of (MLS)-S-2-SVM, so it is necessary for users to identify proper parameters in advance. Originality/value On the basis of MTLS-SVM, a novel multi-label classification approach, (MLS)-S-2-SVM, is put forward. This method can overcome the unbalanced problem but also explicitly models arbitrary order correlations among labels by allowing multiple labels to share a subspace. In addition, the multi-label classification approach has a wider range of applications. That is to say, it is not limited to the field of image classification.

Sequential Labeling with Structural SVM under Nondecomposable Losses.

Sequential Labeling with Structural Svm under the F-1 Loss

Sequential Labeling with Structural SVM under an Average Precision Loss.

A Novel Svm Modeling Approach For Highly Imbalanced And Overlapping Classification

Labelwise Margin Maximization For Sequence Labeling

Parallel and sequential support vector machines for multi-label classification

Sequence Multi-Labeling: A Unified Video Annotation Scheme with Spatial and Temporal Context

Structural SVM with Partial Ranking for Activity Segmentation and Classification

Exploring Image Specific Structured Loss For Image Annotation With Incomplete Labelling

Diversified Hidden Markov Models for Sequential Labeling

An Efficient Decomposition Framework for Discriminative Segmentation with Supermodular Losses

Sequential Label Enhancement.

Sequential Data Classification in the Space of Liquid State Machines.

An improved MLTSVM using label-specific features with missing labels

Image Annotation with Incomplete Labelling by Modelling Image Specific Structured Loss

Labeling Sequential Data Based on Word Representations and Conditional Random Fields

Modeling sequential annotations for sequence labeling with crowds

Incomplete Multi-View Partial Multi-Label Classification Via Deep Semantic Structure Preservation

ML2S-SVM: multi-label least-squares support vector machine classifiers.

OptSLA: an Optimization-Based Approach for Sequential Label Aggregation