Technical Report, SFB 475: Komplexitätsreduktion in Multivariaten Datenstrukturen, Universität Dortmund 1999,50
Most of the literature on combination of forecasts deals with the assumption of unbiased individual forecasts. Here, we consider the case of biased forecasts and discuss two different combination techniques resulting in an unbiased forecast. On the one hand we correct the individual forecasts, and on the other we calculate bias based weights. A simulation study gives some insight in the situations where we should use the different methods.
combination of forecasts bias correction regression generalized Jackknife multivariate forecasts