Introduction: ECAI 2016 is the best place for European AI scientific achievements. The conference provides researchers with a good opportunity to introduce and listen to the best research results of contemporary artificial intelligence.
A Scalable Clustering-Based Local Multi-Label Classification Method
Summary: The goal of multi-label classification is to assign multiple tags to a single test instance. Recently, more and more applications of multi-label classification have a large-scale problem, in which the number of instances, features, and tags is either very large or very large. To solve this problem, in this paper, we develop a cluster-based local multi-label classification method that attempts to reduce the size of instances, features, and labels. Our method consists of low dimensional data clustering and local model learning. Collectively speaking, by applying cluster analysis to subspaces of features, the original data set is first decomposed into several regular-scale parts. This process is induced by supervised multi-label reduction techniques; then, on each data cluster, Training an efficient local multi-label model, META tag classifier chain. Given a test instance, only the local model belonging to the most recent cluster is activated for prediction. Extensive experiments on eighteen benchmark datasets have demonstrated that the proposed method is as efficient as the best algorithm.
First author introduction
Lu Sun
Position; PhD, Graduate School of Information Science and Technology, Hokkaido University
Research direction: Large-scale multi-label classification, multi-label feature selection, clustering high-dimensional data
Related academic papers:
· "Multi-Label Classification with Meta-Label-Specific Features" (ICPR2016)
· "Fast Random k-labelsets for Large-Scale Multi-Label Classification" (ICPR2016)
Via:ECAI 2016
PS : This article was compiled by Lei Feng Network (search "Lei Feng Network" public number) and it was compiled without permission.
Original paper download
