Gradient-Based Optimization of Hyperparameters for Base-Pairing Profile Local Alignment Kernels


Kengo Sato[1][2][3] (sato-kengo@aist.go.jp)
Yutaka Saito[3] (saito@dna.bio.keio.ac.jp)
Yasubumi Sakakibara[3] (yasu@bio.keio.ac.jp)

[1] Japan Biological Informatics Consortium (JBIC), 2-45 Aomi, Koto-ku, Tokyo 135-8073, Japan
[2] Computational Biology Research Center (CBRC), National Institute of Advanced Industrial Science and Technology (AIST), 2-42 Aomi, Koto-ku, Tokyo 135-0064, Japan
[3] Department of Biosciences and Informatics, Keio University, 3-14-1 Hiyoshi, Kohoku-ku, Yokohama, Kanagawa 223-8522, Japan

Abstract

We have recently proposed novel kernel functions, called base-pairing profile local alignment (BPLA) kernels for discrimination and detection of functional RNA sequences using SVMs. We employ STRAL's scoring function which takes into account sequence similarities as well as upstream and downstream base-pairing probabilities, which enables us to model secondary structures of RNA sequences. In this paper, we develop a method for optimizing hyperparameters of BPLA kernels with respect to discrimination accuracy using a gradient-based optimization technique. Our experiments show that the proposed method can find a nearly optimal set of parameters much faster than the grid search on all parameter combinations.

[ Full-text PDF |Table of Contents ]


Japanese Society for Bioinformatics