2023, Vol. 8, Issue 5, Part A
Evaluation of classification ability of support vector machine (SVM) in binary classification problems
Author(s): Halagundegowda GR, Abhishek Singh, Mohan Kumar TL and Naveena K
Abstract: The objective of the study is to investigate the classification ability of SVM architecture including internal parameters and kernel types on farmers' classification based on drought coping strategies. Support vector machines with sigmoid kernel trick, having hyper parameter gamma=0.096 and classification type 1 with capacity C =2.40. The optimal value of hyper parameter has been computed by i=1000 number of iterations for tuning the model by random grid search optimization approach with sigmoid kernel trick and coefficient value 0.06. SVM with the Sigmoid kernel trick has 97 overall support vectors and 93 bounded support vectors. The classification summary of SVM Sigmoid depicts that overall, 93.3% of these cases were correctly classified by the model and the remaining 7% were wrongly classified. The SVM with RBF kernel trick, having hyperparam gamma=0.083 and classification type 1 with capacity C =1.20. The optimal value of the hyperparameter has been computed by i=1000 number of iterations for tuning the model by random grid search optimization approach with radial basis function kernel trick. SVM with RBF kernel trick has 82 overall support vectors and 74 bounded support vectors.
Pages: 07-13 | Views: 152 | Downloads: 17Download Full Article: Click Here
How to cite this article:
Halagundegowda GR, Abhishek Singh, Mohan Kumar TL, Naveena K. Evaluation of classification ability of support vector machine (SVM) in binary classification problems. Int J Stat Appl Math 2023;8(5):07-13.
