Performance comparison of different kernel functions in SVM for different k value in k-fold cross-validation
Author:
- 15 October 2020
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Abstract
It is very important to classify datasets with high accuracy in order to make meaningful decisions in today's science world. In particular, machine learning-based models are known to effectively classify complex datasets. One of the powerful ways to test the success rate of models used for classification is k-fold cross validation. However, very few studies have investigated how k value affects classification results. In this study, the performance of different kernel functions such as Gaussian, linear, and polynomials in SVM were compared for different k values in three different data sets. The most accurate results were obtained with the Gaussian and linear kernel functions.Keywords
This publication has 6 references indexed in Scilit:
- RBF kernel based SVM classification for landmine detection and discrimination2017
- A comparative analysis of SVM and its stacking with other classification algorithm for intrusion detection2016
- Cross Validation Through Two-Dimensional Solution Surface for Cost-Sensitive SVMIEEE Transactions on Pattern Analysis and Machine Intelligence, 2016
- Neural network and SVM classifiers accurately predict lipid binding proteins, irrespective of sequence homologyJournal of Theoretical Biology, 2014
- Review and performance comparison of SVM- and ELM-based classifiersNeurocomputing, 2014
- SVM kernel functions for classification2013