Support Vector Machine Multiclass using Polynomial Kernel for Osteoporosis Detection

Deasy Alfiah Adyanti, Dian Candra Rini Novitasari, Aris Fanani

2018

Abstract

Support Vector Machine is a good performance machine learning algorithm applied as a classification method. In SVM, several problems are difficult to be separated linearly, for mapping data from the lower dimensional to the higher dimensional space, the kernel method is needed. The purpose of this research is to classify data into normal bone, osteopenia, and osteoporosis using SVM Multiclass with polynomial kernel parameters. The classification is based on the analysis of the mandibular ramus bone observed from changes in the trabecular pattern of the jaw bone and hip fracture using SVM Multiclass. Before using SVM multiclass, an image enhancement was performed with adaptive histogram equalization, and feature extraction with the gray level co-occurrence matrix (GLCM). The variable input used in this research is dental panoramic radiograph data as much as 61 data divided into two parts that are 75:25 as training and as test data. Based on the implementation of SVM Multiclass with the polynomial kernel as the basis of computer-aided diagnosis system for osteoporosis detection, the best test data accuracy is 81.25%.

Download


Paper Citation


in Harvard Style

Adyanti D., Novitasari D. and Fanani A. (2018). Support Vector Machine Multiclass using Polynomial Kernel for Osteoporosis Detection.In Proceedings of the International Conference on Mathematics and Islam - Volume 1: ICMIs, ISBN 978-989-758-407-7, pages 384-390. DOI: 10.5220/0008522303840390


in Bibtex Style

@conference{icmis18,
author={Deasy Alfiah Adyanti and Dian Candra Rini Novitasari and Aris Fanani},
title={Support Vector Machine Multiclass using Polynomial Kernel for Osteoporosis Detection},
booktitle={Proceedings of the International Conference on Mathematics and Islam - Volume 1: ICMIs,},
year={2018},
pages={384-390},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0008522303840390},
isbn={978-989-758-407-7},
}


in EndNote Style

TY - CONF

JO - Proceedings of the International Conference on Mathematics and Islam - Volume 1: ICMIs,
TI - Support Vector Machine Multiclass using Polynomial Kernel for Osteoporosis Detection
SN - 978-989-758-407-7
AU - Adyanti D.
AU - Novitasari D.
AU - Fanani A.
PY - 2018
SP - 384
EP - 390
DO - 10.5220/0008522303840390