[Retracted] Constructing a Predictive Model of Depression in Chemotherapy Patients with Non-Hodgkin’s Lymphoma to Improve Medical Staffs’ Psychiatric Care

Figure 4

Validation of the prediction model. (a) Calibration curve of the prediction model of depressive symptomatology in patients with NHL (). The -axis is the predicted risk of depressive symptomatology in patients with non-Hodgkin’s lymphoma. The -axis represents the actual incidence of depressive symptomatology in patients with non-Hodgkin’s lymphoma. The solid line represents the actual prediction of the prediction model. The more consistent the solid line and the dotted line are, the better the predictive ability of the predictive model. (b) AUC represents the accurate prediction of the risk of depressive symptomatology in patients with NHL in randomly selected cases. The AUC values of the training set (red) and validation set (blue) are 0.972 and 0.856, respectively.

(a)

(b)