PLS Regression Using Spatial Weights on the Example of Spatial Modeling Support for Political Parties in Elections 2011 to the Sejm of the Republic of Poland

Maciej Beręsewicz

doi:10.18778/0208-6018.292.13

Authors

Maciej Beręsewicz Poznań University of Economics, Department of Statistics

DOI:

https://doi.org/10.18778/0208-6018.292.13

Abstract

Space has an important role in the reality around us, especially in the context of socioeconomic research. One of the best examples in which the geographic location of one of the most significant factors is the support for political parties. Interesting from the standpoint of policy research is to analyze factors influencing the results of the political party in a particular spatial or administrative unit. The article focuses on the analysis of electoral data for counties.

This was motivated by the high availability of data at a county level, which may be obtained from the Local Data Bank. However, collinearity which occurs in data that affect the support of political parties, limits the use of ordinary linear models. It results in failure of taking into account most of the information contained in the data. In the article will be presented Spatial Partial Least Squares Regression (SPLSR) which takes into account the spatial factor and collinearity.

Author will assess SPLSR model with known spatial linear models with spatial lag and error to compare fit, information criteria and errors. Aim of the article is to show, if taking into account collinearity of predictors significantly improve modelling the support for political parties, which SPLSR model does.

Downloads

Download data is not yet available.

References

Anderson G, Kaufmann P. and Renberg L. (1996), Non-linear modelling with a coupled neural network-PLS regression system. J. Chemom. 10.
Google Scholar DOI: https://doi.org/10.1002/(SICI)1099-128X(199609)10:5/6<605::AID-CEM449>3.0.CO;2-3

Hoskuldsson A. (1988), PLS regression methods. J. Chemom. 2.
Google Scholar DOI: https://doi.org/10.1002/cem.1180020306

Lorber A., Wangen L. and Kowalski B.R. (1987), A theoretical foundation for PLS, J. of Chemometrics 1.
Google Scholar DOI: https://doi.org/10.1002/cem.1180010105

Kopczewska K. (2006), Ekonometria i statystyka przestrzenna z wykorzystaniem programu R CRAN, CeDeWu, Warszawa.
Google Scholar

Suchecki B. (ed.) (2010), Ekonometria Przestrzenna Metody i modele analizy danych, Warszawa, C.H. Beck.
Google Scholar

Wold, H. (1966), Estimation of principal components and related models by iterative least squares, In P.R. Krishnaiaah (Ed.). Multivariate Analysis, (pp. 391-420) New York: Academic Press.
Google Scholar

Wold H. (1981), The fix-point approach to interdependent systems, Amsterdam: North Holland.
Google Scholar DOI: https://doi.org/10.1016/B978-0-444-85451-3.50007-8

Wold H. (1985), Partial least squares, (in:) Samuel Kotz and Norman L. Johnson (eds.), Encyclopedia of statistical sciences, Vol. 6, New York: Wiley, pp. 581-591.
Google Scholar