The Path Kernel: A Novel Kernel for Sequential Data

Andrea Baisero, Florian T. Pokorny, Danica Kragic, Carl Henrik Ek
In Pattern Recognition Applications and Methods, 2015, pp. 71-84

Abstract

We define a novel kernel function for finite sequences of arbitrary length which we call the path kernel. We evaluate this kernel in a classification scenario using synthetic data sequences and show that our kernel can outperform state of the art sequential similarity measures. Furthermore, we find that, in our experiments, a clustering of data based on the path kernel results in much improved interpretability of such clusters compared to alternative approaches such as dynamic time warping or the global alignment kernel.

Files

Download this publication

Bibtex

@incollection{baisero2015a, year={2015}, isbn={978-3-319-12609-8}, booktitle={Pattern Recognition Applications and Methods}, volume={318}, series={Advances in Intelligent Systems and Computing}, editor={Fred, Ana and De Marsico, Maria}, doi={10.1007/978-3-319-12610-4_5}, title={The Path Kernel: A Novel Kernel for Sequential Data}, url={http://link.springer.com/chapter/10.1007/978-3-319-12610-4_5}, publisher={Springer International Publishing}, keywords={Kernels; Sequences}, author={Baisero, Andrea and Pokorny, Florian T. and Kragic, Danica and Ek, Carl Henrik}, pages={71-84}, language={English} }