Label-based multiple kernel learning for classification
Bing Yang, Qian Li, Changhe Fu, Lujia Song, Ling Jing · 2013
This paper provides a novel technique for multiple kernel learning within Support Vector Machine framework. The problem of combining different sources of information arises in several situations, for instance, the classification of data with asymmetric similarity matrices or the construction of an optimal classifier from a collection of kernels. Often, each source of information can be expressed as a similarity matrix. In this paper we propose a new method in order to produce a single optimal kernel matrix from a collection of kernel (similarity) matrices with the label information for classification purposes. Then, the constructed kernel matrix is used to train a Support Vector Machine. The key ideas within the kernel construction are twofold: the quantification, relative to the classification labels, of the difference of information among the similarities; and the linear combination of similarity matrices to the concept of functional combination of similarity matrices. The proposed method has been successfully evaluated and compared with other powerful classifiers on a variety of real classification problems.