Feature selection for classification using decision tree

Nooritawati Md Tahir, Aini Hussain, Salina Abdul Samad, Khairul Anuar Ishak, Rosmawati Abdul Halim

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

17 Citations (Scopus)

Abstract

In most pattern recognition (PR) system, selecting the best feature vectors is an important task. Feature vectors serve as a reduced representation of the original data/signal/input that helps avoid the curse of dimensionality in a PR task. In this work, we consider further effort in selecting the best feature vectors for the PR task. A framework to determine the best eigenvectors of two main human postures based on the rules of thumb of Principal Component Analysis (PCA) has been developed, which includes the KG-rule, Cumulative Variance and the Scree test. Accordingly, two rules of thumb suggest in retaining only the first six eigenvectors or also known as 'eigenposture', as inputs for classification. Using Decision Tree (DT) as our classifier, two distinct processes need to be implemented namely, building the tree and then performing classification using the top-down approach in DT construction. At each node, decision to what is likely to be the best split is determined using the predicted value. Each branch in the tree is labeled with its decision rule whilst each terminal node is labeled with the predicted value of that node. Cross validation ensures selection of an optimum tree size and helps avoid the problem of over fitting. Consequently, the framework has enabled us to select the best and most optimized eigenpostures for classification purpose.

Original languageEnglish
Title of host publicationSCOReD 2006 - Proceedings of 2006 4th Student Conference on Research and Development "Towards Enhancing Research Excellence in the Region"
Pages99-102
Number of pages4
DOIs
Publication statusPublished - 2006
Event2006 4th Student Conference on Research and Development "Towards Enhancing Research Excellence in the Region", SCOReD 2006 - Shah Alam
Duration: 27 Jun 200628 Jun 2006

Other

Other2006 4th Student Conference on Research and Development "Towards Enhancing Research Excellence in the Region", SCOReD 2006
CityShah Alam
Period27/6/0628/6/06

Fingerprint

Decision trees
Feature extraction
Eigenvalues and eigenfunctions
Pattern recognition
Pattern recognition systems
Principal component analysis
Classifiers

Keywords

  • Classification
  • Decision tree
  • Human posture
  • Principal component analysis

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Cite this

Tahir, N. M., Hussain, A., Abdul Samad, S., Ishak, K. A., & Halim, R. A. (2006). Feature selection for classification using decision tree. In SCOReD 2006 - Proceedings of 2006 4th Student Conference on Research and Development "Towards Enhancing Research Excellence in the Region" (pp. 99-102). [4339317] https://doi.org/10.1109/SCORED.2006.4339317

Feature selection for classification using decision tree. / Tahir, Nooritawati Md; Hussain, Aini; Abdul Samad, Salina; Ishak, Khairul Anuar; Halim, Rosmawati Abdul.

SCOReD 2006 - Proceedings of 2006 4th Student Conference on Research and Development "Towards Enhancing Research Excellence in the Region". 2006. p. 99-102 4339317.

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

Tahir, NM, Hussain, A, Abdul Samad, S, Ishak, KA & Halim, RA 2006, Feature selection for classification using decision tree. in SCOReD 2006 - Proceedings of 2006 4th Student Conference on Research and Development "Towards Enhancing Research Excellence in the Region"., 4339317, pp. 99-102, 2006 4th Student Conference on Research and Development "Towards Enhancing Research Excellence in the Region", SCOReD 2006, Shah Alam, 27/6/06. https://doi.org/10.1109/SCORED.2006.4339317
Tahir NM, Hussain A, Abdul Samad S, Ishak KA, Halim RA. Feature selection for classification using decision tree. In SCOReD 2006 - Proceedings of 2006 4th Student Conference on Research and Development "Towards Enhancing Research Excellence in the Region". 2006. p. 99-102. 4339317 https://doi.org/10.1109/SCORED.2006.4339317
Tahir, Nooritawati Md ; Hussain, Aini ; Abdul Samad, Salina ; Ishak, Khairul Anuar ; Halim, Rosmawati Abdul. / Feature selection for classification using decision tree. SCOReD 2006 - Proceedings of 2006 4th Student Conference on Research and Development "Towards Enhancing Research Excellence in the Region". 2006. pp. 99-102
@inproceedings{29211f62dfe748b7951aa31c02c82a65,
title = "Feature selection for classification using decision tree",
abstract = "In most pattern recognition (PR) system, selecting the best feature vectors is an important task. Feature vectors serve as a reduced representation of the original data/signal/input that helps avoid the curse of dimensionality in a PR task. In this work, we consider further effort in selecting the best feature vectors for the PR task. A framework to determine the best eigenvectors of two main human postures based on the rules of thumb of Principal Component Analysis (PCA) has been developed, which includes the KG-rule, Cumulative Variance and the Scree test. Accordingly, two rules of thumb suggest in retaining only the first six eigenvectors or also known as 'eigenposture', as inputs for classification. Using Decision Tree (DT) as our classifier, two distinct processes need to be implemented namely, building the tree and then performing classification using the top-down approach in DT construction. At each node, decision to what is likely to be the best split is determined using the predicted value. Each branch in the tree is labeled with its decision rule whilst each terminal node is labeled with the predicted value of that node. Cross validation ensures selection of an optimum tree size and helps avoid the problem of over fitting. Consequently, the framework has enabled us to select the best and most optimized eigenpostures for classification purpose.",
keywords = "Classification, Decision tree, Human posture, Principal component analysis",
author = "Tahir, {Nooritawati Md} and Aini Hussain and {Abdul Samad}, Salina and Ishak, {Khairul Anuar} and Halim, {Rosmawati Abdul}",
year = "2006",
doi = "10.1109/SCORED.2006.4339317",
language = "English",
isbn = "1424405270",
pages = "99--102",
booktitle = "SCOReD 2006 - Proceedings of 2006 4th Student Conference on Research and Development {"}Towards Enhancing Research Excellence in the Region{"}",

}

TY - GEN

T1 - Feature selection for classification using decision tree

AU - Tahir, Nooritawati Md

AU - Hussain, Aini

AU - Abdul Samad, Salina

AU - Ishak, Khairul Anuar

AU - Halim, Rosmawati Abdul

PY - 2006

Y1 - 2006

N2 - In most pattern recognition (PR) system, selecting the best feature vectors is an important task. Feature vectors serve as a reduced representation of the original data/signal/input that helps avoid the curse of dimensionality in a PR task. In this work, we consider further effort in selecting the best feature vectors for the PR task. A framework to determine the best eigenvectors of two main human postures based on the rules of thumb of Principal Component Analysis (PCA) has been developed, which includes the KG-rule, Cumulative Variance and the Scree test. Accordingly, two rules of thumb suggest in retaining only the first six eigenvectors or also known as 'eigenposture', as inputs for classification. Using Decision Tree (DT) as our classifier, two distinct processes need to be implemented namely, building the tree and then performing classification using the top-down approach in DT construction. At each node, decision to what is likely to be the best split is determined using the predicted value. Each branch in the tree is labeled with its decision rule whilst each terminal node is labeled with the predicted value of that node. Cross validation ensures selection of an optimum tree size and helps avoid the problem of over fitting. Consequently, the framework has enabled us to select the best and most optimized eigenpostures for classification purpose.

AB - In most pattern recognition (PR) system, selecting the best feature vectors is an important task. Feature vectors serve as a reduced representation of the original data/signal/input that helps avoid the curse of dimensionality in a PR task. In this work, we consider further effort in selecting the best feature vectors for the PR task. A framework to determine the best eigenvectors of two main human postures based on the rules of thumb of Principal Component Analysis (PCA) has been developed, which includes the KG-rule, Cumulative Variance and the Scree test. Accordingly, two rules of thumb suggest in retaining only the first six eigenvectors or also known as 'eigenposture', as inputs for classification. Using Decision Tree (DT) as our classifier, two distinct processes need to be implemented namely, building the tree and then performing classification using the top-down approach in DT construction. At each node, decision to what is likely to be the best split is determined using the predicted value. Each branch in the tree is labeled with its decision rule whilst each terminal node is labeled with the predicted value of that node. Cross validation ensures selection of an optimum tree size and helps avoid the problem of over fitting. Consequently, the framework has enabled us to select the best and most optimized eigenpostures for classification purpose.

KW - Classification

KW - Decision tree

KW - Human posture

KW - Principal component analysis

UR - http://www.scopus.com/inward/record.url?scp=46849106073&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=46849106073&partnerID=8YFLogxK

U2 - 10.1109/SCORED.2006.4339317

DO - 10.1109/SCORED.2006.4339317

M3 - Conference contribution

AN - SCOPUS:46849106073

SN - 1424405270

SN - 9781424405275

SP - 99

EP - 102

BT - SCOReD 2006 - Proceedings of 2006 4th Student Conference on Research and Development "Towards Enhancing Research Excellence in the Region"

ER -