Table 1 Performance evaluation of selected models based on proper scoring rules
From: Understanding norovirus reporting patterns in England: a mixed model approach
Model | RPS | logS | ||
|---|---|---|---|---|
Endemic seasonality (S = 2) + linear trend + covariates | Mean score | Calibration test (p-Value) | Mean score | Calibration test (p-Value) |
No random effects: | ||||
A1 epidemic (S = 0) | 0.938 | 0.385 | 1.480 | 0.270 |
A2 epidemic (S = 1) | 0.934 | 0.251 | 1.480 | 0.177 |
Uncorrelated random effects: | ||||
B1 epidemic (S = 0) | 0.890 | 0.888 | 1.430 | 0.499 |
B2 epidemic (S = 1) | 0.884 | 0.514 | 1.430 | 0.245 |
Correlated random effects: | ||||
C1 epidemic (S = 0) | 0.890 | 0.810 | 1.430 | 0.448 |
C2 epidemic (S = 1) | 0.884 | 0.452 | 1.430 | 0.211 |