Fig. 5 | Scientific Reports

Fig. 5

From: Integrating single-cell RNA sequencing, WGCNA, and machine learning to identify key biomarkers in hepatocellular carcinoma

Fig. 5

Identification of hub genes via the LASSO model, the SVM-RFF model, and the RF model. A Venn diagram illustrating the overlapping genes (A). The correlation between the number of trees in the random forest and the error rates (B). The top 25 significant genes were identified from the random forest. MeanDecreaseGini shows the rank of genes based on their relative importance (C). The error of 10-fold cross-validation (CV) in SVM-RFE algorithms (D). The plot of LASSO coefficient profiles (E). Each curve represents a single gene. The plot of partial likelihood deviance (F). Vertical dashed lines were plotted at the optimal lambda values. The intersection of key genes from three machine learning (G). The ROC curves and AUC values of the hub genes (H).

Back to article page