Premium
A New Method of Regression of Time‐series Based on a Mathematical Connection Between Path and Compartmental Analyses
Author(s) -
Eöry A.
Publication year - 1981
Publication title -
biometrical journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.108
H-Index - 63
eISSN - 1521-4036
pISSN - 0323-3847
DOI - 10.1002/bimj.4710230706
Subject(s) - series (stratigraphy) , path (computing) , path coefficient , time series , mathematics , regression analysis , regression , connection (principal bundle) , path analysis (statistics) , statistics , analogy , ordinary least squares , econometrics , computer science , paleontology , linguistics , philosophy , geometry , biology , programming language
In the regression analysis of time‐series the error terms may be serially correlated with the results. In this case the possibility arises that the expectations and adjustment processes are themselves mis‐specifications of the correct behavioural relationships. In this paper an analogy is pointed out between the compartmental‐analysis and the path analysis by the author. It will be argued that the time‐series of the standardized partial regression coefficients (path coefficients or beta weights) computed from the cross‐sectional data, have to be analysed instead of applying ordinary least squares directly to the timeseries. An application of the new method is briefly discussed.