Skip to main content

Table 8 The best regression model found by the MCMC sampler

From: Estimating the size of urban populations using Landsat images: a case study of Bo, Sierra Leone, West Africa

(1) Covariates (2) Estimated coefficient (3) Std. Error (4) t value (5) Pr(\(>|t|\)) (6) Max. Prob. (7) VIF (8) PIP
(intercept) − 174.02 20.16 − 8.63 9.63E−7 < .0001
nb7v − 1656.91 85.58 − 19.36 5.72E−11 < .0001 6.29 0.9837
r_sp37 532.23 31.27 17.02 2.88E−10 < .0001 1.34 0.9790
nb1v 1686.14 71.28 23.66 4.52E−12 < .0001 5.36 0.9500
r_sp15s − 2744.29 159.21 − 17.24 2.46E−10 < .0001 1.08 0.4711
ch245c 44775.32 2636.73 16.98 2.96E−10 < .0001 2.72 0.9835
r_sp14c − 246.75 23.79 − 10.37 1.18E−07 < .0001 1.55 0.5381
  1. This table summarizes the best regression equation returned by the MCMC sampler for the estimation of \(\sqrt{d}\). The values of the variance inflation factor (VIF) are less than 7.0, which demonstrates the low collinearity between the covariates. Four of Posterior Inclusion Probabilities (PIPs) are close to 1.0, quantifying their importance as predictive variables of \(\sqrt{d}\), as discussed in the text