Graph kernels based on tree patterns for molecules

DOI10.1007/S10994-008-5086-2MaRDI QIDQ1959635zbMATH OpenOpenAlexWikidataFDO

Authors Pierre Mahé, Jean-Philippe Vert

Publication date 7 October 2010

Published in Machine Learning (Search for Journal in Brave)

Copyright license Creative Commons Attribution-NonCommercial 4.0 International

Full work available at URL https://arxiv.org/abs/q-bio/0609024

zbMATH Keywords

support vector machinesgraph kernelschemoinformatics

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Graph theory (including graph drawing) in computer science (68R10) Computational methods for problems pertaining to biology (92-08)

Abstract: Motivated by chemical applications, we revisit and extend a family of positive definite kernels for graphs based on the detection of common subtrees, initially proposed by Ramon et al. (2003). We propose new kernels with a parameter to control the complexity of the subtrees used as features to represent the graphs. This parameter allows to smoothly interpolate between classical graph kernels based on the count of common walks, on the one hand, and kernels that emphasize the detection of large common subtrees, on the other hand. We also propose two modular extensions to this formulation. The first extension increases the number of subtrees that define the feature space, and the second one removes noisy features from the graph representations. We validate experimentally these new kernels on binary classification tasks consisting in discriminating toxic and non-toxic molecules with support vector machines.

Recommendations

• Two new graph kernels and applications to chemoinformatics
• Graph kernels
• Treelet kernel incorporating chiral information
• Treelet kernel incorporating cyclic, stereo and inter pattern information in chemoinformatics
• Weisfeiler-Lehman graph kernels

Cites work

• Complete mining of frequent patterns from graphs: Mining graph data
• On graph kernels: hardness results and efficient alternatives.

Cited in

(16)

• The weight function in the subtree kernel is decisive
• Effective graph classification based on topological and label attributes
• Propagation kernels: efficient graph kernels from propagated information
• Realizable cycle structures in digraphs
• Incremental embedding within a dissimilarity-based framework
• Deep graph similarity learning: a survey
• The journey of graph kernels through two decades
• Kernel discriminant analysis and clustering with parsimonious Gaussian process models
• Treelet kernel incorporating cyclic, stereo and inter pattern information in chemoinformatics
• Two new graph kernels and applications to chemoinformatics
• Learning molecular classes from small numbers of positive examples using graph grammars
• Relevant cycle hypergraph representation for molecules
• Treelet kernel incorporating chiral information
• Graph kernels: a survey
• A long trip in the charming world of graphs for pattern recognition
• Evolutionary Construction of Granular Kernel Trees for Cyclooxygenase-2 Inhibitor Activity Comparison

Describes a project that uses

Uses Software

• LIBSVM

This page was built for publication: Graph kernels based on tree patterns for molecules

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1959635)