Semi-Parametric Tools for Spatial Hedonic Models: An Introduction to Mixed Geographically Weighted Regression and Geoadditive Models

  • Ghislain Geniaux
  • Claude Napoléone
This chapter focuses on the contribution of semi-parametric tools such as Mixed Geographically Weighted Regression (MGWR) (Fotheringham et al. 1997) and Geoadditive Models (Hastie and Tibshirani 1993; Kammann and Wand 2003) that belong to the General Additive Models family (GAM), to explore the effects of distance regressors on price and/or spatial nonstationarity of implicit house price coefficient in hedonic functions. These tools have become essential with:
  • Growing evidence in the economic literature of the multiplicity of amenity sources linked to space and location of houses. Amenity sources can be related to the distance to different public services, commercial facilities, noteworthy landmarks, tourist area, etc. Using monocentric models based on the single distance to Central Business District (CBD) is not sufficient to assess the effect of urban sprawl on land prices. Dominant models in economy fail to account for amenity values, congestion cost or social expenditure for public infrastructure (Brueckner 2001). The specification of hedonic OLS models may be quite difficult when numerous distance measures are candidate regressors, especially with respect to the colinearity issue. GAM is a powerful and flexible tool to explore distance effects with large samples.

  • The upscale and downscale extension of available spatial descriptors provided by many public and private organizations has resulted in an explosion in the use of GIS in the last five years. This increased use of GIS by public administrations facilitates the introduction of finer descriptors of landscape, public services and social neighbourhoods in hedonic models. Working on large samples with a lot of finer descriptors of the environment and neighborhood involves dealing with numerous local effects. In this case it is necessary to use multiple scales to analyze the structure of house prices (Quigley 1995). Locally weighted regressions such as Geographically Weighted Regressions (GWR) or MGWR, and GAM, enable estimation of spatially varying and invariant implicit prices in hedonic analysis.


Spatial Autocorrelation House Price Generalize Additive Model Geographically Weight Regression Central Business District 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Anselin L (1988) Spatial Econometrics: Methods and Models. Kluwer Academic Publishers, Dordrecht, The Netherlands.Kluwer Academic Publishers, Dordrecht, The NetherlandsGoogle Scholar
  2. Anselin L (2001) Spatial effects in econometric practice in environmental and resource economics. American Journal of Agricultural Economics 83: 705- 710CrossRefGoogle Scholar
  3. Anselin L, Bera AK, Florax R, Yoon MJ (1996) Simple diagnostic test for spatial dependance. Regional Science and Urban Economics 26: 77-104CrossRefGoogle Scholar
  4. Azzalini A, Bowman A (1997) Applied Smoothing Techniques for Data Analysis. Oxford University Press, NewYorkGoogle Scholar
  5. Bivand R, Yu D (2006) Geographically weighted regression. R package version 0.3-12,
  6. Bowman A (1984) An alternative method of cross-validation for the smoothing of density estimate. Biometrika 71: 353-360CrossRefGoogle Scholar
  7. Brueckner JK (2001) Urban sprawl: Lessons from urban economics. In: Gale WG, Rothenberg J (eds) Brookings-Wharton Papers on Urban Affairs. Brookings Institution Press, pp 65-89Google Scholar
  8. Brunsdon C, Fotheringham AS, Charlton M (1996) Geographically weighted regression: A method for exploring spatial nonstationarity. Geographical Analysis 28: 281-298Google Scholar
  9. Brunsdon C, Fotheringham AS, Charlton M (1998) Spatial nonstationarity and autoregressive models. Environment and Planning A 30: 957-993CrossRefGoogle Scholar
  10. Cleveland W (1979) Robust locally weighted regression and smoothing scatter-plots. Journal of the American Statistical Association 74: 829-836CrossRefGoogle Scholar
  11. Craven P, Wahba G (1979) Smoothing noisy data with spline functions. Numerische Mathematik 31: 377-403CrossRefGoogle Scholar
  12. Des Rosiers F, Thiérault M (1996) Houses prices, neighborhood factors and spatial dynamics: improving hedonic modeling using trend surface analysis. Working Paper, Faculté des sciences de l’Administration, Laval University, Québec, CAGoogle Scholar
  13. Fotheringham AS, Charlton M, Brunsdon C (1997) Two techniques for exploring nonstationarity in geographical data. Geographical Systems 4: 59-82Google Scholar
  14. Fotheringham AS, Brunsdon C, Charlton M (1999) Some notes on parametric signi¿cance tests for geographically weigthed regression. Journal of Regional Science 39: 497-524CrossRefGoogle Scholar
  15. Fotheringham AS, Brunsdon C, Charlton M (2002) Geographically Weighted Regression: The Analysis of Spatially Varying Relationships. John Wiley & Sons, New YorkGoogle Scholar
  16. Geniaux G, Napoléone C (2005) Rente foncière et anticipations dans le périurbain. Economie et Prévision 68: 77-92Google Scholar
  17. Hastie TJ, Tibshirani R (1990) Generalized Additive Models. Chapman and Hall, LondonGoogle Scholar
  18. Hastie TJ, Tibshirani R (1993) Varying-coef¿cient models. Journal of the Royal Statistical Society B55: 757-796Google Scholar
  19. Hoaglin D, Welsch RE (1978) The hat matrix in regression and anova. The American Statistician 32: 17-22CrossRefGoogle Scholar
  20. Kammann EE, Wand MP (2003) Geoadditive models. Journal of the Royal Statistical Society C52: 1-18Google Scholar
  21. Kendall M, Stuart A (1977) The advanced theory of statistics. Grif¿n, LondonGoogle Scholar
  22. Kim YJ, Gu C (2004) Smoothing spline gaussian regression: more scalable computation via ef¿cient approximation. Journal of the Royal Statistical Society B66: 337-356Google Scholar
  23. Leung Y, Mei C, Zhang W (2000a) Statistical tests for spatial nonstationarity based on the geographically weighted regression model. Environment And Planning A32: 9-32CrossRefGoogle Scholar
  24. Leung Y, Mei C, Zhang W (2000b) Testing for spatial autocorrelation among the residuals of the geographically weighted regression. Environment And Planning A32: 871-890CrossRefGoogle Scholar
  25. Mei C (2006) Geographically weighted regression technique for spatial data analysis. Working Paper of School Science, Xi’an Jiaotong UniversityGoogle Scholar
  26. Mei C, He S, Fang K (2004) A note on the mixed geographically weighted regression model. Journal of Regional Science 44: 143-157CrossRefGoogle Scholar
  27. Paez A (2002) Spatial parametric non-stationarity: a variance heterogeneity approach. Working Paper, Centre for Spatial Analysis, School of Geography and Geology, McMaster UniversityGoogle Scholar
  28. Paez A, Uchida T, Miyamoto K (2001) Spatial association and heterogeneity issues in land price models. Urban Studies 38: 1493-1508CrossRefGoogle Scholar
  29. Plantinga A, Lubowski R, Stavins R (2002) The effects of potential land development on agricultural land prices. Journal of Urban Economics 52: 561-581CrossRefGoogle Scholar
  30. Quigley JM (1995) A simple hybrid model for estimating real estate price indexes. Journal of Housing Economics 4: 1-12CrossRefGoogle Scholar
  31. Speckman P (1988) Kernel smoothing in partial linear model. Journal of the Royal Statistical Society B50: 413-436Google Scholar
  32. Wood S (2004) Stable and efficient multiple smoothing parameter estimation for generalized additive models. Journal of the American Statistical Association 99: 673-686CrossRefGoogle Scholar
  33. Wood S (2006) Generalized Additive Models: An Introduction with R. Chapman and Hall, CRCGoogle Scholar

Copyright information

© Springer Science + Business Media, LLC 2008

Authors and Affiliations

  • Ghislain Geniaux
    • 1
  • Claude Napoléone
    • 1
  1. 1.INRA EcodeveloppementAvignonFrance

Personalised recommendations