Differentiation of the Follicular Neoplasm on the Gray-Scale US by Image Selection Subsampling along with the Marginal Outline Using Convolutional Neural Network

Table 8

Result of various typical inference model.

	Predicted_Adenoma	Predicted_Carcinoma		Overall Accuracy
SVM	True_Adenoma; 18.30%	False_Carcinoma; 81.70%	True negative rate; 0.183	40.96%
	False_Adenoma; 22.92%	True_Carcinoma; 77.08%	True positive rate; 0.7708
	False omission rate; 0.4400	Positive predictive value; 0.3718	-score: 0.4148
			-score: 0.5017
			-score: 0.6346
			-mean: 0.5354
KNN	True_Adenoma; 91.50%	False_Carcinoma; 8.50%	True negative rate; 0.9150	63.45%
	False_Adenoma; 81.25%	True_Carcinoma; 18.75%	True positive rate; 0.1875
	False omission rate; 0.3578	Positive predictive value; 0.5806	-score: 0.4091
			-score: 0.2835
			-score: 0.2169
			-mean: 0.3299
ANN	True_Adenoma; 79.08%	False_Carcinoma; 20.92%	True negative rate; 0.7908	70.28%
	False_Adenoma; 43.75%	True_Carcinoma; 56.25%	True positive rate; 0.5625
	False omission rate; 0.2577	Positive predictive value; 0.6279	-score: 0.6136
			-score: 0.5934
			-score: 0.5745
			-mean: 0.5943
Normal Bayes Classifier	True_Adenoma; 38.56%	False_Carcinoma; 61.44%	True negative rate; 0.3856	57.03%
	False_Adenoma; 13.54%	True_Carcinoma; 86.46%	True positive rate; 0.8646
	False omission rate; 0.1805	Positive predictive value; 0.4689	-score: 0.5162
			-score: 0.6081
			-score: 0.7398
			-mean: 0.6367