What is it about?
In forecasting a variable, reference class forecasting uses past forecasts of comparable variables. More traditional statistical modeling using forecasts based on developed a model of the variable of interest. This paper shows how both kinds of forecasts can be integrated into a combined forecast superior to both the reference class forecast and the model-based forecast.
Featured Image
Why is it important?
Accurate forecasts are essential to a wide variety of practical contexts. Model-based forecasts are known to have certain limitations which has motivated the introduction of reference class forecasting as an alternative.
Perspectives
In systems engineering, there is a major debate over the value of model-based forecasts given some widely publicized failures. Some have argued for abandoning model-based forecasts quickly. This paper argues for a more balanced solution.
Dr Robert F Bordley
University of Michigan, Ann Arbor
Read the Original
This page is a summary of: Reference Class Forecasting: Resolving Its Challenge to Statistical Modeling, The American Statistician, October 2014, Taylor & Francis,
DOI: 10.1080/00031305.2014.937544.
You can read the full text:
Contributors
The following have contributed to this page
