Performance of the Hybrid Approach Using Three Machine Learning Algorithms

Betul Kan-Kilinc; Yonca Yazirli

doi:10.18187/pjsor.v16i2.3069

No Detail Found!

Performance of the Hybrid Approach Using Three Machine Learning Algorithms

Author(s)

Betul Kan-Kilinc Eskisehir Technical University, Faculty of Science, Department of Statistics, Turkey

Yonca Yazirli Eskisehir Technical University, Institute of Graduate Programs, Department of Statistics,Turkey

Abstract

One of the essential problems in data mining is the removal of negligible variables from the data set. This paper proposes a hybrid approach that uses rough set theory based algorithms to reduct the attribute selected from the data set and utilize reducts to raise the classification success of three learning methods; multinomial logistic regression, support vector machines and random forest using 5-fold cross validation. The performance of the hybrid approach is measured by related statistics. The results show that the hybrid approach is effective as its improved accuracy by 6-12% for the three learning methods.

Publication Details

Page(s) 217-224

DOI 10.18187/pjsor.v16i2.3069

Published Journal: Pakistan Journal of Statistics and Operation Research, Volume: 16, Issue: 2, Year: 2020

Request the full text from PASTIC

Keywords

performance, Reduction, Rough set, Accuracy.

Citations

Downloads

Views