Skip to main content

Table 2 Best models selected from fitting to the first nine seasons and the corresponding prediction error obtained from predictions for the last three seasons of the OOH GPC data

From: Influenza epidemic surveillance and prediction based on electronic health record data from an out-of-hours general practitioner cooperative: model development and validation on 2003–2015 data

Epidemic

Endemic

k

logSa

WAICa

MSEb

 

AR(1)

β 0 + β 1 t + S t  + C s

5

64.8761

1101.893

13.8822

M1

AR(1)

β 0 + β 1 t + S t

 

64.9161

1103.298

17.8292

M2

AR(1)

β 0 + S t  + C s

5

64.9217

1101.101

14.5707

M3

AR(1)

β 0 + S t  + C s

4

64.8822

1104.608

17.1564

M4

AR(1)

β 0 + S t

 

65.1286

1101.176

16.6048

M5

RW-1

β 1 t + S t  + C s

5

64.5683

1103.741

21.1512

M6

AR(1)

β 1 t + S t

 

64.6006

1103.310

20458.71

M7

AR(1)

S t  + C s

5

64.5406

1102.243

14.2192

M8

IID

S t

 

403.3084

984.414

17.1342

M9

AR(1)

S t

 

64.3999

1102.030

908.6617

M10

AR(1)

β 0 + β 1 t + sin(2πt/52) + cos(2πt/52)

1.4243

1108.511

22.7856

M11

  1. All the presented models used Poisson likelihood for the ILI counts. An extended table can be found [http://goo.gl/n5kHbU]; aUsed the first nine seasons data, bfor the last three seasons data; logS logarithmic score, WAIC Watanabe-Akaike information criteria, MSE mean square error