Figure 6 | Scientific Reports

Figure 6

From: Exploration of effective biomarkers for venous thrombosis embolism in Behçet’s disease based on comprehensive bioinformatics analysis

Figure 6

Identification of candidate diagnostic biomarkers through machine learning algorithms. (A) The Lasso regression algorithm identified six highly accurate biomarkers for diagnosing VTE with BD by precisely locating the curve's lowest point. (B, C) The random forest algorithm was employed to assess the error in VTE, leading to the ranking of the top 15 genes based on their importance scores. (D, E) The SVM-RFE algorithm selected 18 genes with the lowest error and highest accuracy. (F) Four hub genes (E2F1, GATA3, HDAC5, and MSH2) were identified as candidate biomarkers based on the intersection of the results from the three algorithms.

Back to article page