A 10-Gene Signature Identified by Machine Learning for Predicting the Response to Transarterial Chemoembolization in Patients with Hepatocellular Carcinoma

Table 1

Comparative performance between different predictive models using selective genes.

Models	AUC	F1 score	Accuracy	Youden index	Sensitivity	Specificity	PPV	NPV
SVM	0.930	0.798	0.823	0.643	0.808	0.835	0.793	0.851
ANN	0.922	0.816	0.843	0.690	0.815	0.865	0.832	0.867
Log	0.918	0.819	0.847	0.686	0.815	0.871	0.830	0.865
XGBoost	0.858	0.762	0.793	0.581	0.769	0.812	0.763	0.827
RF	0.839	0.726	0.760	0.515	0.738	0.776	0.737	0.804

Log: Lasso-logistic regression; SVM: support vector machine; ANN: artificial neural network; RF: random forest; XGBoost: eXtreme gradient boosting-based tree model.