Skip to main content
. 2013 Aug 26;8(1):42–52. doi: 10.1111/irv.12158

Table 6.

Multivariate time-series models for ILI and SARI (ILI: November 2008–2009; n = 1·321 and SARI: November 2008–2009; n = 845)

β SE P value
ILI
 ARIMA (1,1,3) – ARCH (2) model
  AR1 −0·968 0·024 <0·001
  MA1 0·421 0·056 <0·001
  MA2 −0·638 0·046 <0·001
  MA3 −0·172 0·049 <0·001
  A0 12·509 0·282 <0·001
  A1 0·218 0·082 0·007
  A2 0·535 0·074 <0·001
 Sunshine duration −0·395 0·211 0·03
  AIC 2264·2
  BIC 2287·6
  Jarque–Bera test <0·001
  Ljung–Box test 0·973
SARI
 ARIMA (0,1,3) – ARCH (4) model
  MA1 −0·703 0·053 <0·001
  MA2 −0·176 0·059 0·003
  MA3 0·111 0·049 0·026
  A0 1·551 0·269 <0·001
  A1 0·131 0·087 0·067
  A2 0·395 0·103 <0·001
  A3 0·241 0·078 <0·001
  A4 0·158 0·078 0·021
 Mean temperature 0·399 0·209 0·025
 Sunshine duration −0·392 0·160 0·007
 Relative humidity −0·098 0·059 0·05
 Mean concentration of pollutants −0·079 0·037 0·018
  AIC 1857·2
  BIC 1888·3
Jarque–Bera test 0·014
Ljung–Box test 0·949

ILI, influenza-like illness; SARI, severe acute respiratory infection; ARIMA, autoregressive integrated moving average; ARCH, autoregressive conditional heteroskedasticity; AIC, Akaike information criterion; BIC, Bayesian information criterion.