Fig. 4. A diagnosis and severity prediction model for COVID-19 patients. (A) The diagnostic model for asymptomatic and presymptomatic patients. The Receiver Operating Characteristic (ROC) curve of diagnostics of training (red line) and five-fold cross-validation (dotted lines). The dotted lines in different colors represent the results of 5-fold cross-validation. (B) The contribution degrees of markers (Top 10) in the diagnostic model. The horizontal axis represents the contribution degree of laboratory indicators. (C) The prognostic model for the severity of presymptomatic patients. (D) The contribution degrees of markers (Top 10) in the prognostic model. The horizontal axis represents the contribution degree of laboratory indicators. (E) and (F) A comparison of laboratory indicators between non-severe and severe patients without symptom on admission. The data set in (E) comes from Wuhan Huoshenshan Hospital in China. The validation data set in (F) comes from the First People’s Hospital of Jiangxia District of Wuhan, Hubei, China. ** represents P -value < 0.01. * represents P -value < 0.05.