Fig. 4

Identifying potential biomarkers for CD. A The SVM-RFE model. The horizontal axis denotes the number of featured genes, while the vertical axis represents the error rate in curve variation after cross-validation. In the graph, N = 4 in the figure indicates that there are 4 feature genes with the lowest error rate, which is close to zero. B The LASSO model. The horizontal axis displays the logarithmic punishment coefficient, log λ, while the vertical axis shows the error of the cross-validation. A lower value on the Y-axis indicates a better fitting result of the equation. The two dashed lines indicate two specific lambda (λ) values. The dashed line on the left represents λ min, which indicates the lambda value when the bias is minimal, signifying that the model fitting is the best at this lambda value. In this study, λ min on the left was chosen as the final criterion for selecting the equation. The dashed line on the right represents λ-se, which refers to one standard error to the right of the minimum λ value; C in colon tissue, LASSO and SVM-RFE share biomarkers