Tree-Based Kernel for Graphs With Continuous Attributes

Giovanni Martino, Nicolo Navarin, Alessandro Sperduti

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

The availability of graph data with node attributes that can be either discrete or real-valued is constantly increasing. While existing Kernel methods are effective techniques for dealing with graphs having discrete node labels, their adaptation to nondiscrete or continuous node attributes has been limited, mainly for computational issues. Recently, a few kernels especially tailored for this domain, and that trade predictive performance for computational efficiency, have been proposed. In this brief, we propose a graph kernel for complex and continuous nodes' attributes, whose features are tree structures extracted from specific graph visits. The kernel manages to keep the same complexity of the state-of-the-art kernels while implicitly using a larger feature space. We further present an approximated variant of the kernel, which reduces its complexity significantly. Experimental results obtained on six real-world data sets show that the kernel is the best performing one on most of them. Moreover, in most cases, the approximated version reaches comparable performances to the current state-of-the-art kernels in terms of classification accuracy while greatly shortening the running times.

Original languageEnglish
JournalIEEE Transactions on Neural Networks and Learning Systems
DOIs
Publication statusAccepted/In press - 13 Jun 2017

Fingerprint

Computational efficiency
Labels
Availability

Keywords

  • Big data applications
  • Bioinformatics
  • Computational complexity
  • Feature extraction
  • Kernel
  • Learning systems
  • machine learning
  • Proteins
  • supervised learning
  • support vector machines.

ASJC Scopus subject areas

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

Cite this

Tree-Based Kernel for Graphs With Continuous Attributes. / Martino, Giovanni; Navarin, Nicolo; Sperduti, Alessandro.

In: IEEE Transactions on Neural Networks and Learning Systems, 13.06.2017.

Research output: Contribution to journalArticle

@article{a6e20d73d4d149129121838902ecb8f5,
title = "Tree-Based Kernel for Graphs With Continuous Attributes",
abstract = "The availability of graph data with node attributes that can be either discrete or real-valued is constantly increasing. While existing Kernel methods are effective techniques for dealing with graphs having discrete node labels, their adaptation to nondiscrete or continuous node attributes has been limited, mainly for computational issues. Recently, a few kernels especially tailored for this domain, and that trade predictive performance for computational efficiency, have been proposed. In this brief, we propose a graph kernel for complex and continuous nodes' attributes, whose features are tree structures extracted from specific graph visits. The kernel manages to keep the same complexity of the state-of-the-art kernels while implicitly using a larger feature space. We further present an approximated variant of the kernel, which reduces its complexity significantly. Experimental results obtained on six real-world data sets show that the kernel is the best performing one on most of them. Moreover, in most cases, the approximated version reaches comparable performances to the current state-of-the-art kernels in terms of classification accuracy while greatly shortening the running times.",
keywords = "Big data applications, Bioinformatics, Computational complexity, Feature extraction, Kernel, Learning systems, machine learning, Proteins, supervised learning, support vector machines.",
author = "Giovanni Martino and Nicolo Navarin and Alessandro Sperduti",
year = "2017",
month = "6",
day = "13",
doi = "10.1109/TNNLS.2017.2705694",
language = "English",
journal = "IEEE Transactions on Neural Networks and Learning Systems",
issn = "2162-237X",
publisher = "IEEE Computational Intelligence Society",

}

TY - JOUR

T1 - Tree-Based Kernel for Graphs With Continuous Attributes

AU - Martino, Giovanni

AU - Navarin, Nicolo

AU - Sperduti, Alessandro

PY - 2017/6/13

Y1 - 2017/6/13

N2 - The availability of graph data with node attributes that can be either discrete or real-valued is constantly increasing. While existing Kernel methods are effective techniques for dealing with graphs having discrete node labels, their adaptation to nondiscrete or continuous node attributes has been limited, mainly for computational issues. Recently, a few kernels especially tailored for this domain, and that trade predictive performance for computational efficiency, have been proposed. In this brief, we propose a graph kernel for complex and continuous nodes' attributes, whose features are tree structures extracted from specific graph visits. The kernel manages to keep the same complexity of the state-of-the-art kernels while implicitly using a larger feature space. We further present an approximated variant of the kernel, which reduces its complexity significantly. Experimental results obtained on six real-world data sets show that the kernel is the best performing one on most of them. Moreover, in most cases, the approximated version reaches comparable performances to the current state-of-the-art kernels in terms of classification accuracy while greatly shortening the running times.

AB - The availability of graph data with node attributes that can be either discrete or real-valued is constantly increasing. While existing Kernel methods are effective techniques for dealing with graphs having discrete node labels, their adaptation to nondiscrete or continuous node attributes has been limited, mainly for computational issues. Recently, a few kernels especially tailored for this domain, and that trade predictive performance for computational efficiency, have been proposed. In this brief, we propose a graph kernel for complex and continuous nodes' attributes, whose features are tree structures extracted from specific graph visits. The kernel manages to keep the same complexity of the state-of-the-art kernels while implicitly using a larger feature space. We further present an approximated variant of the kernel, which reduces its complexity significantly. Experimental results obtained on six real-world data sets show that the kernel is the best performing one on most of them. Moreover, in most cases, the approximated version reaches comparable performances to the current state-of-the-art kernels in terms of classification accuracy while greatly shortening the running times.

KW - Big data applications

KW - Bioinformatics

KW - Computational complexity

KW - Feature extraction

KW - Kernel

KW - Learning systems

KW - machine learning

KW - Proteins

KW - supervised learning

KW - support vector machines.

UR - http://www.scopus.com/inward/record.url?scp=85021790026&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85021790026&partnerID=8YFLogxK

U2 - 10.1109/TNNLS.2017.2705694

DO - 10.1109/TNNLS.2017.2705694

M3 - Article

AN - SCOPUS:85021790026

JO - IEEE Transactions on Neural Networks and Learning Systems

JF - IEEE Transactions on Neural Networks and Learning Systems

SN - 2162-237X

ER -