MRI-based brain tumor detection using convolutional deep learning methods and chosen machine learning techniques

Table 10 Multiple comparisons between methods

Dependent variable	(I) method	(J) method	Mean difference (I–J)	SE	Sig	95% Confidence interval
Dependent variable	(I) method	(J) method	Mean difference (I–J)	SE	Sig	Lower bound	Upper bound
Precision	2D CNN	LR	.32750*	.11921	.011	.0815	.5735
		SGD	.40000*	.11921	.003	.1540	.6460
		MLP	.58250*	.11921	.000	.3365	.8285
	Convolutional auto-encoder	LR	.32250*	.11921	.012	.0765	.5685
		SGD	.39500*	.11921	.003	.1490	.6410
		MLP	.57750*	.11921	.000	.3315	.8235
Recall	2D CNN	LR	.33500*	.12520	.013	.0766	.5934
		SGD	.40750*	.12520	.003	.1491	.6659
		MLP	.70500*	.12520	.000	.4466	.9634
	Convolutional auto-encoder	LR	.32000*	.12520	.017	.0616	.5784
		SGD	.39250*	.12520	.004	.1341	.6509
		MLP	.69000*	.12520	.000	.4316	.9484
F-measure	2D CNN	SVM	.14500*	.06854	.045	.0035	.2865
		LR	.33000*	.06854	.000	.1885	.4715
		SGD	.44500*	.06854	.000	.3035	.5865
		MLP	.79000*	.06854	.000	.6485	.9315
	Convolutional auto-encoder	LR	.32000*	.06854	.000	.1785	.4615
		SGD	.43500*	.06854	.000	.2935	.5765
		MLP	.78000*	.06854	.000	.6385	.9215

ISSN: 1472-6947