Skip to main content

Advertisement

Table 2 Summary characteristics from the trend SARerr model selection of BT spread in France in 2007-2008, based on clinical cases (n = 10 994 municipalities).

From: Estimating front-wave velocity of infectious diseases: a simple, efficient method applied to bluetongue

model np AIC minRSA R 2 observed Moran's I
TSA 6 115 121 26.5 0.85 0.3894
m0 8 107 730 5.9 0.92 0.0285
m1 7 107 740 5.9 0.92 0.0298
m2 7 107 750 6.0 0.92 0.0319
m3 7 107 750 6.0 0.92 0.0309
m4 7 107 740 5.9 0.92 0.0299
m5 7 107 730 5.9 0.96 0.0282
m6 5 107 780 6.0 0.92 0.0345
  1. Model selection was based on Akaike Information Criterion (AIC) and minimum residual spatial autocorrelation (minRSA). A measure of model fit and spatial autocorrelation in the model residuals are given as R2; and observed Moran's I, respectively. np is the number of parameters. m0 to m6 are the trend SARerr models, TSA model values are given for comparison. The selected trend SARerr model is m5.
  2. TSA: t = β0 + β1X + β2Y + β3X2 + β4XY + β5Y2 + ε
  3. m0: t = β0 + β1X + β2Y + β3X2 + β4XY + β5Y2 + λWμ + ε
  4. m1: t = β0 + β1X + β2Y + β3X2 + β4XY + λWμ + ε
  5. m2: t = β0 + β1X + β2Y + β4XY + β5Y2 + λWμ + ε
  6. m3: t = β0 + β1X + β2Y + β3X2 + β5Y2 + λWμ + ε
  7. m4: t = β0 + β1X + β3X2 + β4XY + β5Y2 + λWμ + ε
  8. m5: t = β0 + β2Y + β3X2 + β4XY + β5Y2 + λWμ + ε
  9. m6: t = β0 + β1X + β2Y + λWμ + ε.