Skip to main content

Table 2 Comparison of predictive accuracy of three models across years for each country using multivariate Diebold-Mariano (DM) test

From: A multi-country comparison of stochastic models of breast cancer mortality with P-splines smoothing approach

Country Null Hypothesis Test Statistic Value P-Value Decision
China H0: SLC=FDM = BMS
OR
H0: All model hold Equal predictive accuracy
− 1796.2 0.000 H0 is rejected
Pakistan −20.151 0.000 H0 is rejected
India −24.314 0.000 H0 is rejected
Thailand −3.9457 0.000 H0 is rejected
  1. Predictive period, 1990–2010