Table 4 Results using other Document Embedding and machine learning models.

Document Embedding	Model	Accuracy	Sensitivity	Specificity
Original algorithm	DNN	0.900	0.881	0.916
	GNB	0.817	0.674	0.933
	LR	0.831	0.674	0.958
	SVC	0.863	0.756	0.95
	XGboost	0.829	0.731	0.908
TF-IDF	DNN	0.824	0.798	0.845
	GNB	0.78	0.617	0.912
	LR	0.752	0.482	0.971
	SVC	0.785	0.565	0.962
	XGboost	0.785	0.653	0.891
BERT	DNN	0.847	0.762	0.916
	GNB	0.833	0.762	0.891
	LR	0.447	1	0
	SVC	0.845	0.710	0.95
	XGboost	0.826	0.731	0.904

TF-IDF term frequency–inverse document frequency, BERT bidirectional encoder representations from transformers, DNN deep neural network, GNB Gaussian Naive Bayes, LR logistic regression, SVC support vector machine classifier.

Quick links

Search