Noninvasive Evaluation of Portal Hypertension Using a Supervised Learning Technique

Table 6

Comparison of classification results.

Attributes	Acc, %	Sn	Sp	F1	AUC	Classifier
Demographic, laboratory data, liver/spleen TE. Our method	89.72	0.83	0.92	0.87	0.96	Meta voting
Measuring hyaluronan and laminin in serum [17]	—	0.83	0.82	0.83	—	Logistic regression
Albumin, INR, ALT [32]	77.00	0.93	0.37	0.53	—	Predictive model
Demographic, laboratory data [19]	—	0.82	0.76	0.79	0.82	Regression
Serum, liver TE [18]	80.00	0.82	0.83	0.82	0.84	Neural network
Liver TE [14]	88.70	0.88	0.88	0.88	0.95	Statistical