Minimization of the mean square error of the preliminary estimates of the Canadian beneficiaries series / by Guy Huot. : CS11-614/93-4E-PDF

"There are sometimes systematic biases in preliminary observations. The statistical agencies producing the figures make adjustments to obtain unbiased preliminary observations. These figures are first published and then revised one or two months later. The revised figures are the true values. This paper uses the structure of the data revisions to obtain unbiased preliminary observations. The results encompass the in-sample analysis that support the monthly release by Statistics Canada of the unbiased preliminary observations of ten Canadian unemployment insurance beneficiaries series, and a summary of the performance of the out-of-sample released values from January 88 to December 91. The unbiased preliminary observations have reduced level and direction errors due to bias, in 95.4% of the 480 cases studied in the out-of-sample analysis"--Abstract.

Lien permanent pour cette publication :
publications.gc.ca/pub?id=9.838754&sl=1

• Format MARC XML
• Format MARC HTML

Demander des formats alternatifs