Table 6.

Three classes performance comparison of the proposed method and (Chandra et al., 2021) that is using majority voting (ground truth and mathematical evaluation).

	Chandra et al. (2020)	Proposed Method	Accuracy (%)
Phase-1	Naïve Bayes	Cosine KNN	88.372	92.2
	Decision Tree	Linear Discriminant	90.698	92.4
	KNN	Bagged Trees Ensemble	95.736	95
	SVM (Poly Kernel)	SqueezeNet Deep Learning	96.124	95.86
	ANN	Medium Gaussian SVM	96.512	97.2
	Majority Voting (Ground Truth)		98.062	–
	Majority Voting (Mathematical Evaluation)		99.79	99.86
Phase-2	KNN	Logistic Regression	72.093	87.5
	ANN	Linear Discriminant	73.256	89.5
	Decision Tree	Bagged Trees Ensemble	79.070	90.6
	Naïve Bayes	Cosine KNN	80.814	91.7
	SVM (RBF Kernel)	Medium Gaussian SVM	86.628	93.3
	Majority Voting (Ground Truth)		91.329	–
	Majority Voting (Mathematical Evaluation)		93.08	99.28
Overall	Majority Voting (Ground Truth)		93.41	–
	Majority Voting (Mathematical Evaluation)		97.11	99.63