Premium
An Appraisal of Composite Forecasting Methods
Author(s) -
Park David W.,
Tomek William G.
Publication year - 1988
Publication title -
applied economic perspectives and policy
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.4
H-Index - 49
eISSN - 2040-5804
pISSN - 2040-5790
DOI - 10.1093/aepp/10.1.1
Subject(s) - monte carlo method , commodity , econometrics , composite number , selection (genetic algorithm) , computer science , operations research , economics , engineering , machine learning , statistics , mathematics , algorithm , finance
This paper assesses the potential benefits of composite forecasting methods as applied to commodity prices. Conditions under which composite forecasts are most applicable are examined. A Monte Carlo simulation and empirical examples are used to assess the benefits of composite forecasts and provide guidelines on the selection of appropriate methods. The importance of independent pieces of information and why composites cannot be a cure for poor forecasts under many realistic situations are discussed.