*
* Example 2/4 from pp 41-50
*
open data beer2.dat
calendar(m) 1991
data(format=free,org=columns) 1991:1 1995:8 beer
*
data(format=free,unit=input) 1995:1 1995:8 forecast
 150.25 139.50 157.25 143.50 138.00 127.50 138.25 141.50 140.50 167.25
*
* The uforeerrors procedure does the calculation of forecast error statistics.
* Its standard syntax is
*
* @uforeerrors actual forecast start end
*
* You can skip the "start end" parameters if the only place where the two series
* overlap are the entries you want to evaluate.
*
@uforeerrors beer forecast
*
* Note, the Theil's Relative U (which is described on pp 48-49 of the text, is
* being reported slightly differently here because uforeerrors uses all eight
* forecasts in the calculation, not seven.