Construction and Validation of Predictive Model to Identify Critical Genes Associated with Advanced Kidney Disease

Figure 5

The model simulation. (a) Principal Component Analysis (PCA) results. PC1 and PC2, respectively, represent the two principal components with the largest variance value in PCA results. The points with different colors in the figures represent the Disease and Normal samples, respectively. (b) ROC curves analysis for the comparison of five training models. FPR represents False Positive Rate , and TPR refers to True Positive Rate . Various colors represented different training models. A larger area under ROC curves indicated a better prediction feature of models. (c) The top 20 features were evaluated by the XGB model. score indicates the feature importance. A larger score refers to the more importance of the gene in the XGB model.

(a)

(b)

(c)