Hybrid Information Gain Method and Bagging in Data Classification using Support Vector Machine

Immanuel H. G. Manurung, Tulus, Poltak Sihombing

2018

Abstract

The selection process is very influential attributes of dataset in SVM algorithm which tends to produce good accuracy on the results of the classification (classifier) ​​SVM is not optimal. To reduce the effect of selection attributes on SVM classifiers, it is necessary to apply a combination of methods of feature selection algorithms that are Bootstrapping Aggregation (Bagging) and methods of Information Gain (IG). The application of the algorithm Bagging the feature selection is done to give weight to each feature are recommended, so that the found feature is a strong classifier, whereas IG focuses on identifying attributes and evaluate the impact of a beneficial features based on ranking the features that can be recommended to the classifier SVM in the process classification. Experiments implementation of Information Gain feature selection techniques that use attributes with election threshold level. The results showed that, the performance accuracy of SVM classifiers in dataset by combining IG before bagging process, by setting the value thresold >= 0.02 and a 10-fold cross-validation, show that with the implementation of information gain feature selection techniques can improve the performance of machine learning classification algorithm.

Paper Citation

in Harvard Style

Manurung I., Tulus. and Sihombing P. (2018). Hybrid Information Gain Method and Bagging in Data Classification using Support Vector Machine. In Proceedings of the 3rd International Conference of Computer, Environment, Agriculture, Social Science, Health Science, Engineering and Technology - Volume 1: ICEST, ISBN 978-989-758-496-1, pages 194-201. DOI: 10.5220/0010040401940201

in Bibtex Style

@conference{icest18,
author={Immanuel H. G. Manurung and Tulus and Poltak Sihombing},
title={Hybrid Information Gain Method and Bagging in Data Classification using Support Vector Machine},
booktitle={Proceedings of the 3rd International Conference of Computer, Environment, Agriculture, Social Science, Health Science, Engineering and Technology - Volume 1: ICEST,},
year={2018},
pages={194-201},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010040401940201},
isbn={978-989-758-496-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 3rd International Conference of Computer, Environment, Agriculture, Social Science, Health Science, Engineering and Technology - Volume 1: ICEST,
TI - Hybrid Information Gain Method and Bagging in Data Classification using Support Vector Machine
SN - 978-989-758-496-1
AU - Manurung I.
AU - Tulus.
AU - Sihombing P.
PY - 2018
SP - 194
EP - 201
DO - 10.5220/0010040401940201