ON LIMITATIONS TO PREDICTION

Thomas A. Gleeson Florida State University

Search for other papers by Thomas A. Gleeson in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

An expression is obtained for the limit to certainty of perfect prediction, in terms of the amount of verification data and the detail of the predictand. Related expressions are developed for the number of effective predictors that may be combined in a given prediction method, and for the fineness of predictand detail that can be forecast.

Applications to evaluation of practical forecast problems are discussed by use of examples. Similar treatments are suggested for specification problems.

Because of the generality of their derivations, it is believed that these expressions would be useful in other fields, as well as in meteorology.

Abstract

An expression is obtained for the limit to certainty of perfect prediction, in terms of the amount of verification data and the detail of the predictand. Related expressions are developed for the number of effective predictors that may be combined in a given prediction method, and for the fineness of predictand detail that can be forecast.

Applications to evaluation of practical forecast problems are discussed by use of examples. Similar treatments are suggested for specification problems.

Because of the generality of their derivations, it is believed that these expressions would be useful in other fields, as well as in meteorology.

Save