Combining neural networks and decision trees

Ananth Sankar, Richard J. Mammone

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Neural networks and decision trees are two common approaches to pattern recognition. In this paper, these approaches are combined to develop a new neural network architecture based on decision trees and a new learning rule to grow this architecture using neural network techniques. The resulting neural network is called a neural tree network (NTN). The NTN can be implemented very efficiently as compared to multilayer perceptrons (MLP). The learning algorithm is more efficient than the exhaustive search techniques used in standard decision tree methods. The algorithm also grows the network, thus finding the correct number of neurons as opposed to the backpropagation algorithm used to train MLPs in which the number of neurons and their interconnections must be known before learning can begin. Two different approaches are presented to grow the NTN based on self-organizing clustering techniques and a supervised learning rule. Simulation results are presented on a speaker-independent vowel recognition task which show the superiority of the NTN approach over both MLPs and decision trees.

Original languageEnglish (US)
Title of host publicationProceedings of SPIE - The International Society for Optical Engineering
PublisherPubl by Int Soc for Optical Engineering
Pages374-383
Number of pages10
Editionpt 1
ISBN (Print)0819405787
StatePublished - Jan 1 1991
EventApplications of Artificial Neural Networks II - Orlando, FL, USA
Duration: Apr 2 1991Apr 5 1991

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Numberpt 1
Volume1469
ISSN (Print)0277-786X

Other

OtherApplications of Artificial Neural Networks II
CityOrlando, FL, USA
Period4/2/914/5/91

Fingerprint

Decision trees
Neural networks
Neurons
learning
Backpropagation algorithms
Supervised learning
Multilayer neural networks
Network architecture
Learning algorithms
Pattern recognition
neurons
self organizing systems
vowels
organizing
pattern recognition

All Science Journal Classification (ASJC) codes

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering

Cite this

Sankar, A., & Mammone, R. J. (1991). Combining neural networks and decision trees. In Proceedings of SPIE - The International Society for Optical Engineering (pt 1 ed., pp. 374-383). (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 1469, No. pt 1). Publ by Int Soc for Optical Engineering.
Sankar, Ananth ; Mammone, Richard J. / Combining neural networks and decision trees. Proceedings of SPIE - The International Society for Optical Engineering. pt 1. ed. Publ by Int Soc for Optical Engineering, 1991. pp. 374-383 (Proceedings of SPIE - The International Society for Optical Engineering; pt 1).
@inproceedings{9e7b3c3c950e4ef7950bf1a185c5f3d1,
title = "Combining neural networks and decision trees",
abstract = "Neural networks and decision trees are two common approaches to pattern recognition. In this paper, these approaches are combined to develop a new neural network architecture based on decision trees and a new learning rule to grow this architecture using neural network techniques. The resulting neural network is called a neural tree network (NTN). The NTN can be implemented very efficiently as compared to multilayer perceptrons (MLP). The learning algorithm is more efficient than the exhaustive search techniques used in standard decision tree methods. The algorithm also grows the network, thus finding the correct number of neurons as opposed to the backpropagation algorithm used to train MLPs in which the number of neurons and their interconnections must be known before learning can begin. Two different approaches are presented to grow the NTN based on self-organizing clustering techniques and a supervised learning rule. Simulation results are presented on a speaker-independent vowel recognition task which show the superiority of the NTN approach over both MLPs and decision trees.",
author = "Ananth Sankar and Mammone, {Richard J.}",
year = "1991",
month = "1",
day = "1",
language = "English (US)",
isbn = "0819405787",
series = "Proceedings of SPIE - The International Society for Optical Engineering",
publisher = "Publ by Int Soc for Optical Engineering",
number = "pt 1",
pages = "374--383",
booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",
edition = "pt 1",

}

Sankar, A & Mammone, RJ 1991, Combining neural networks and decision trees. in Proceedings of SPIE - The International Society for Optical Engineering. pt 1 edn, Proceedings of SPIE - The International Society for Optical Engineering, no. pt 1, vol. 1469, Publ by Int Soc for Optical Engineering, pp. 374-383, Applications of Artificial Neural Networks II, Orlando, FL, USA, 4/2/91.

Combining neural networks and decision trees. / Sankar, Ananth; Mammone, Richard J.

Proceedings of SPIE - The International Society for Optical Engineering. pt 1. ed. Publ by Int Soc for Optical Engineering, 1991. p. 374-383 (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 1469, No. pt 1).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Combining neural networks and decision trees

AU - Sankar, Ananth

AU - Mammone, Richard J.

PY - 1991/1/1

Y1 - 1991/1/1

N2 - Neural networks and decision trees are two common approaches to pattern recognition. In this paper, these approaches are combined to develop a new neural network architecture based on decision trees and a new learning rule to grow this architecture using neural network techniques. The resulting neural network is called a neural tree network (NTN). The NTN can be implemented very efficiently as compared to multilayer perceptrons (MLP). The learning algorithm is more efficient than the exhaustive search techniques used in standard decision tree methods. The algorithm also grows the network, thus finding the correct number of neurons as opposed to the backpropagation algorithm used to train MLPs in which the number of neurons and their interconnections must be known before learning can begin. Two different approaches are presented to grow the NTN based on self-organizing clustering techniques and a supervised learning rule. Simulation results are presented on a speaker-independent vowel recognition task which show the superiority of the NTN approach over both MLPs and decision trees.

AB - Neural networks and decision trees are two common approaches to pattern recognition. In this paper, these approaches are combined to develop a new neural network architecture based on decision trees and a new learning rule to grow this architecture using neural network techniques. The resulting neural network is called a neural tree network (NTN). The NTN can be implemented very efficiently as compared to multilayer perceptrons (MLP). The learning algorithm is more efficient than the exhaustive search techniques used in standard decision tree methods. The algorithm also grows the network, thus finding the correct number of neurons as opposed to the backpropagation algorithm used to train MLPs in which the number of neurons and their interconnections must be known before learning can begin. Two different approaches are presented to grow the NTN based on self-organizing clustering techniques and a supervised learning rule. Simulation results are presented on a speaker-independent vowel recognition task which show the superiority of the NTN approach over both MLPs and decision trees.

UR - http://www.scopus.com/inward/record.url?scp=0025795028&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0025795028&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0025795028

SN - 0819405787

T3 - Proceedings of SPIE - The International Society for Optical Engineering

SP - 374

EP - 383

BT - Proceedings of SPIE - The International Society for Optical Engineering

PB - Publ by Int Soc for Optical Engineering

ER -

Sankar A, Mammone RJ. Combining neural networks and decision trees. In Proceedings of SPIE - The International Society for Optical Engineering. pt 1 ed. Publ by Int Soc for Optical Engineering. 1991. p. 374-383. (Proceedings of SPIE - The International Society for Optical Engineering; pt 1).