Abstract
Multi-label feature selection has drawn wide attention in recent years. The existing multi-label feature selection algorithms mainly assume that the labels of the training data are obtained before feature selection takes place. However, this assumption does not always founded because the acquisition of labeling data is costly. In real-world applications, the available labels usually arrive one by one over time. To address this problem, we develop a novel multi-label feature selection method under the circumstance of streaming label to select a set of the most relevant and discriminative features. Specifically, we firstly select label-specific features for each newly-arrived label by designing inter-class discrimination and intra-class neighbor recognition. Then, a feature conversion is created to fuse the generated label-specific feature sets. Comprehensive experiments on a series of benchmark data sets clearly demonstrate the superiority of the proposed method against other state-of-the-art multi-label feature selection methods.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have