Multiple Similarity Matrices for Multi-label Learning
Jie Zhang · Journal of Information and Computational Science · 2014
In Multi-label learning, each instance is associated with multiple labels. The assignment is to predict the unknown labels of the instances. In this paper, a multi-label learning approach named ML-MKI is contributed. In ML-MKI, not only the similarity of the instances but also the similarity of the labels is all utilized to predict the labels of the instance more accurately in semi-supervised multi-label classiflcation problem. Besides we give an algorithm to solve the ML-MKI model. Experiment on a real-world multi-label bioinformatic data shows that our algorithm ML-MKI has better performance than existing multi-label learning algorithms.