Skip to main content


Fig. 2 | BMC Infectious Diseases

Fig. 2

From: Spatial epidemiological determinants of severe fever with thrombocytopenia syndrome in Miyazaki, Japan: a GWLR modeling study

Fig. 2

Predicted risk map of severe fever with thrombocytopenia syndrome cases using geographically weighted logistic regression. a Geographically weighted logistic regression (GWLR) model to predict all cases and (b) cases in 2017. All cases during 2013–2017 were used as the learning data of A; data from 2013 to 2016 were used for B. The color changes from green to red with elevated risk. Low-risk areas below the maximum Youden index value are colored green, and otherwise yellow to red. Observed cases in 2017 are empty circles; the remaining orange circles represent cases diagnosed during 2013–2016. Receiver operating characteristic curves for correctly predicting SFTS cases in Miyazaki using (c) the entire dataset and (d) data from 2013 to 2016. In addition to GWLR, the results from the logistic regression model are also overlaid

Back to article page