An efficient topological distance-based tree kernel

Fabio Aiolli, Giovanni Da San Martino, Alessandro Sperduti

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Tree kernels proposed in the literature rarely use information about the relative location of the substructures within a tree. As this type of information is orthogonal to the one commonly exploited by tree kernels, the two can be combined to enhance state-of-the-art accuracy of tree kernels. In this brief, our attention is focused on subtree kernels. We describe an efficient algorithm for injecting positional information into a tree kernel and present ways to enlarge its feature space without affecting its worst case complexity. The experimental results on several benchmark datasets are presented showing that our method is able to reach state-of-the-art performances, obtaining in some cases better performance than computationally more demanding tree kernels.

Original languageEnglish
Article number6844015
Pages (from-to)1115-1120
Number of pages6
JournalIEEE Transactions on Neural Networks and Learning Systems
Volume26
Issue number5
DOIs
Publication statusPublished - 1 May 2015

Fingerprint

Information use

Keywords

  • Kernel methods
  • kernels for structured data
  • learning in structured domains
  • position aware kernels
  • tree kernels.

ASJC Scopus subject areas

  • Software
  • Computer Science Applications
  • Computer Networks and Communications
  • Artificial Intelligence

Cite this

An efficient topological distance-based tree kernel. / Aiolli, Fabio; Da San Martino, Giovanni; Sperduti, Alessandro.

In: IEEE Transactions on Neural Networks and Learning Systems, Vol. 26, No. 5, 6844015, 01.05.2015, p. 1115-1120.

Research output: Contribution to journalArticle

@article{dc0419c084b14cac9a10ffd1b2003b66,
title = "An efficient topological distance-based tree kernel",
abstract = "Tree kernels proposed in the literature rarely use information about the relative location of the substructures within a tree. As this type of information is orthogonal to the one commonly exploited by tree kernels, the two can be combined to enhance state-of-the-art accuracy of tree kernels. In this brief, our attention is focused on subtree kernels. We describe an efficient algorithm for injecting positional information into a tree kernel and present ways to enlarge its feature space without affecting its worst case complexity. The experimental results on several benchmark datasets are presented showing that our method is able to reach state-of-the-art performances, obtaining in some cases better performance than computationally more demanding tree kernels.",
keywords = "Kernel methods, kernels for structured data, learning in structured domains, position aware kernels, tree kernels.",
author = "Fabio Aiolli and {Da San Martino}, Giovanni and Alessandro Sperduti",
year = "2015",
month = "5",
day = "1",
doi = "10.1109/TNNLS.2014.2329331",
language = "English",
volume = "26",
pages = "1115--1120",
journal = "IEEE Transactions on Neural Networks and Learning Systems",
issn = "2162-237X",
publisher = "IEEE Computational Intelligence Society",
number = "5",

}

TY - JOUR

T1 - An efficient topological distance-based tree kernel

AU - Aiolli, Fabio

AU - Da San Martino, Giovanni

AU - Sperduti, Alessandro

PY - 2015/5/1

Y1 - 2015/5/1

N2 - Tree kernels proposed in the literature rarely use information about the relative location of the substructures within a tree. As this type of information is orthogonal to the one commonly exploited by tree kernels, the two can be combined to enhance state-of-the-art accuracy of tree kernels. In this brief, our attention is focused on subtree kernels. We describe an efficient algorithm for injecting positional information into a tree kernel and present ways to enlarge its feature space without affecting its worst case complexity. The experimental results on several benchmark datasets are presented showing that our method is able to reach state-of-the-art performances, obtaining in some cases better performance than computationally more demanding tree kernels.

AB - Tree kernels proposed in the literature rarely use information about the relative location of the substructures within a tree. As this type of information is orthogonal to the one commonly exploited by tree kernels, the two can be combined to enhance state-of-the-art accuracy of tree kernels. In this brief, our attention is focused on subtree kernels. We describe an efficient algorithm for injecting positional information into a tree kernel and present ways to enlarge its feature space without affecting its worst case complexity. The experimental results on several benchmark datasets are presented showing that our method is able to reach state-of-the-art performances, obtaining in some cases better performance than computationally more demanding tree kernels.

KW - Kernel methods

KW - kernels for structured data

KW - learning in structured domains

KW - position aware kernels

KW - tree kernels.

UR - http://www.scopus.com/inward/record.url?scp=85027937173&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85027937173&partnerID=8YFLogxK

U2 - 10.1109/TNNLS.2014.2329331

DO - 10.1109/TNNLS.2014.2329331

M3 - Article

VL - 26

SP - 1115

EP - 1120

JO - IEEE Transactions on Neural Networks and Learning Systems

JF - IEEE Transactions on Neural Networks and Learning Systems

SN - 2162-237X

IS - 5

M1 - 6844015

ER -