Fig. 6From: Development of a 32-gene signature using machine learning for accurate prediction of inflammatory bowel diseaseThe 19-gene subtype classification model could distinguish between UC and CD. A-D Heatmaps show the gene expression of the 19 selected genes in iHMP (A), GSE3365 (B), GSE10616 (C), and GSE6731 (D). E The histogram comparisons of the Accuracy, AUC, Recall, Precision, F1, and Kappa of two XGBoost-based classification models, those values range between 0 (poor performance) and 1 (good performance). F-I Confusion matrix for the 19-gene subtype classification model (F), the 32-gene subtype classification model (G), the 54-gene subtype classification model (H), the 5-gene subtype classification model (I)Back to article page