A regression surprise resolved

In this note we explore the following surprising fact: In regression with trend and seasonality, the prediction risk is constant for all seasons of a new cycle, despite the fact that it increases with time when the seasons are left out. Awareness of this may be useful to both the practicing statistician and to teachers of statistics. The challenge of resolving the issue may also be given to students of statistics as a research project.

MoreLess

Year of publication:	2008-09-22
Authors:	Lillestøl, Jostein ; Andersson, Jonas
Institutions:	Institutt for foretaksøkonomi, Norges Handelshøyskole (NHH)
Subject:	Trend and seasonality \| Prediction risk \| Paradox

More details

Extent:	application/pdf
Series:	Discussion Papers.
Type of publication:	Book / Working Paper
Language:	English
Notes:	The text is part of a series Discussion Papers Number 2008/16 6 pages
Classification:	C25 - Discrete Regression and Qualitative Choice Models ; C35 - Discrete Regression and Qualitative Choice Models
Source:	RePEc - Research Papers in Economics

Persistent link: https://www.econbiz.de/10005190561