A Comparative Analysis on Various Extreme Multi-Label Classification Algorithms

In the field of machine learning, the boom in big data has opened a variety of new research problems due to the availability of the extremely huge online data. Extreme Multi-Label Learning (XML) is the most challenging and popular among them. XML addresses the problem of learning a classifier that can automatically tag a data sample with the most relevant subset of labels from a given large label set. For instance, there are more than a million labels (i.e. categories) on Wikipedia and one may wish to build a classifier that can annotate a new article or web page with a subset of relevant Wikipedia categories. Extreme Multi-Label Learning or specifically classification is a very challenging research problem for the need to simultaneously dealing with massive labels, dimensions, and training points. In this paper, we review various approaches such as Embedding, Tree and One-vs-All methods to handle Extreme Multi-Label classification problems and have compared their performance in extreme settings.