Academic
Publications
Multiple Kernel Learning by Conditional Entropy Minimization

Multiple Kernel Learning by Conditional Entropy Minimization,10.1109/ICMLA.2010.40,Hideitsu Hino,Nima Reyhani,Noboru Murata

Multiple Kernel Learning by Conditional Entropy Minimization   (Citations: 1)
BibTex | RIS | RefWorks Download
Kernel methods have been successfully used in many practical machine learning problems. Choosing a suitable kernel is left to the practitioner. A common way to an automatic selection of optimal kernels is to learn a linear combination of element kernels. In this paper, a novel framework of multiple kernel learning is proposed based on conditional entropy minimization criterion. For the proposed framework, three multiple kernel learning algorithms are derived. The algorithms are experimentally shown to be comparable to or outperform kernel Fisher discriminant analysis and other multiple kernel learning algorithms on benchmark data sets.
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
    • ...parameters[6]. We found that MCEM outperforms the margin­...
    • ...tional entropy minimization (MCEM), which was proposed in [6]...
    • ...As stated in [6], we can interpret the classification as supervised di­...
    • ...Since class-conditiona l entropy is upper bounded as follows[6]:...
    • ...a tuning parameter TJ > O. In the present study, we had adopted the random search algorithm in [6]...

    Tetsuji Ogawaet al. Speaker recognition using multiple kernel learning based on conditiona...

Sort by: