The path kernel: A novel kernel for sequential data
2015 (English)In: Pattern Recognition: Applications and Methods : International Conference, ICPRAM 2013 Barcelona, Spain, February 15–18, 2013 Revised Selected Papers / [ed] Ana Fred, Maria De Marsico, Springer Berlin/Heidelberg, 2015, 71-84 p.Conference paper (Refereed)
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.
Place, publisher, year, edition, pages
Springer Berlin/Heidelberg, 2015. 71-84 p.
, Advances in Intelligent Systems and Computing, ISSN 2194-5357 ; 318
Kernels, Sequences, Pattern recognition, Dynamic time warping, Global alignment, Interpretability, Kernel function, Similarity measure, State of the art, Classification (of information)
Computer and Information Science
IdentifiersURN: urn:nbn:se:kth:diva-167388DOI: 10.1007/978-3-319-12610-4_5ISI: 000364822300005ScopusID: 2-s2.0-84914163004ISBN: 9783319126098OAI: oai:DiVA.org:kth-167388DiVA: diva2:815584
2nd International Conference on Pattern Recognition Applications and Methods, ICPRAM 2013; Barcelona; Spain; 15 February 2013 through 18
QC 201506012015-06-012015-05-222015-12-17Bibliographically approved