Fig 1.
Flowchart of the overall methodology.
Table 1.
CBM data from switchgears for classification of abnormal location.
Fig 2.
ANN structure used in this work.
Fig 3.
Concept of classification by SVM (unfilled elements represent support vectors while filled elements represent training data).
Fig 4.
Flowchart of the proposed GSA-AI technique.
Table 2.
Classification results using ANN based on 70:30 ratio for training and testing data.
Table 3.
Classification results using SVM based on 70:30 ratio for training and testing.
Table 4.
Classification results using ANN based on 4-fold cross validation.
Table 5.
Classification results using SVM based on 4-fold cross validation.
Table 6.
Classification results using GSA-ANN based on 70:30 ratio for training and testing data.
Table 7.
Classification results using GSA-SVM based on 70:30 ratio for training and testing data.
Table 8.
Classification results using GSA-ANN based on 4-fold cross validation.
Table 9.
Classification results using GSA-SVM based on 4-fold cross validation.
Fig 5.
Convergence curve of different algorithms combined with (a) ANN and (b) SVM using 70:30 ratio for training and testing data.
Table 10.
Results comparison between different algorithms.
Table 11.
Statistical test of non-parametric using McNemar’s test for the method proposed.