Leveraging machine learning for predicting acute graft-versus-host disease grades in allogeneic hematopoietic cell transplantation for T-cell prolymphocytic leukaemia

Table 2 Model Performance Comparison with Varying Features for acute GvHD grades 0 to 2 vs 3 and 4

Model	Number of features	Training Balanced accuracy	Testing Balanced accuracy	Testing F1 score	Testing ROC AUC
KNeighborsClassifier	5	0.95	0.56	0.77	0.60
LinearDiscriminantAnalysis	5	069	0.57	0.75	0.60
MLPClassifier	5	0.67	0.57	0.75	0.59
KNeighborsClassifier	6	0.93	0.52	0.72	0.60
LinearDiscriminantAnalysis	6	0.70	0.57	0.73	0.60
MLPClassifier	6	0.85	0.52	0.76	0.55
KNeighborsClassifier	9	0.94	0.52	0.73	0.53
LinearDiscriminantAnalysis	9	0.71	0.58	0.74	0.58
MLPClassifier	9	1.00	0.49	0.78	0.50

ISSN: 1471-2288