Skip to main content

Table 6 Point and interval forecast accuracies among the 17 models and two model-averaged methods in the Japanese national data and the average of 47 sub-national populations for ages between 60 and 100+

From: Model confidence sets and forecast combination: an application to age-specific mortality

   RMSFE Mean interval score
Series Method National data Sub-national data National data Sub-national data
Female 1 0.54 1.11 1.81 3.88
  2 4.20 6.24 7.12 61.70
  3 1.34 1.61 4.91 5.41
  4 1.98 2.17 7.24 7.09
  5 1.20 1.51 3.71 4.50
  6 2.05 2.27 5.95 5.84
  7 0.87 1.27 1.93 2.89
  8 0.33 1.02 0.86 2.40
  9 0.67 1.26 2.10 4.26
  10 0.69 1.27 2.24 4.34
  11 0.69 1.27 2.20 4.33
  12 0.69 1.28 2.22 4.30
  13 0.71 1.21 1.31 2.71
  14 0.72 1.21 1.35 2.70
  15 0.61 1.21 1.32 2.94
  16 0.83 1.25 1.43 3.40
  17 0.81 1.23 1.80 2.73
  18 0.54 1.24 0.87 2.57
  19 0.54 1.22 0.87 2.57
Male 1 0.71 2.55 2.79 9.56
  2 2.27 4.02 7.97 14.57
  3 1.93 3.13 8.17 11.40
  4 2.92 3.76 12.18 13.01
  5 1.78 2.96 6.42 9.45
  6 2.75 3.69 9.72 10.47
  7 1.60 2.84 3.81 7.91
  8 0.65 2.47 1.37 5.77
  9 0.99 3.78 3.68 12.80
  10 1.00 3.76 3.80 12.78
  11 1.02 3.83 3.91 13.05
  12 1.00 3.47 3.63 10.88
  13 0.72 2.51 1.50 5.86
  14 0.78 2.51 1.49 5.87
  15 0.72 2.47 1.98 5.02
  16 0.93 2.47 1.79 7.23
  17 0.76 2.44 1.79 5.49
  18 0.71 2.51 1.36 5.06
  19 0.71 2.50 1.36 5.07
  1. Forecast errors have been multiplied by 100. The smallest overall errors are shown in italics