Mining Structured Data
- 19 January 2010
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Computational Intelligence Magazine
- Vol. 5 (1) , 42-49
Abstract
In many application domains, the amount of available data increased so much that humans need help from automatic computerized methods for extracting relevant information. Moreover, it is becoming more and more common to store data that possess inherently structural or relational characteristics. These types of data are best represented by graphs, which can very naturally represent entities, their attributes, and their relationships to other entities. In this article, we review the state of the art in graph mining, and we present advances in processing trees and graphs by two computational intelligence classes of methods, namely neural networks and kernel methods.Keywords
This publication has 39 references indexed in Scilit:
- Graph self-organizing maps for cyclic and unbounded graphsNeurocomputing, 2009
- Graph kernels based on tree patterns for moleculesMachine Learning, 2008
- A Spectrum Tree KernelTransactions of the Japanese Society for Artificial Intelligence, 2007
- Graph miningACM Computing Surveys, 2006
- Universal Approximation Capability of Cascade Correlation for StructuresNeural Computation, 2005
- Online Learning with KernelsIEEE Transactions on Signal Processing, 2004
- A general framework for unsupervised processing of structured dataNeurocomputing, 2004
- A self-organizing map for adaptive processing of structured dataIEEE Transactions on Neural Networks, 2003
- A Discriminative Framework for Detecting Remote Protein HomologiesJournal of Computational Biology, 2000
- Supervised neural networks for the classification of structuresIEEE Transactions on Neural Networks, 1997