Fuzzy Rough Decision Trees for Multi-label Classification

Multi-label classification exists widely in medical analysis or image annotation. Although there are some algorithms to train models for multi-label classification, few of them are able to extract comprehensible rules. In this paper, we propose a multi-label decision tree algorithm based on fuzzy rough sets, named ML-FRDT. This method can tackle with symbolic, continuous and fuzzy data. We conduct experiments on two multi-label datasets. And the experiment results show that ML-FRDT achieves good performance than some well-established multi-label classification algorithms.