Nonparametric regression: Difference between revisions

Revision as of 02:23, 6 January 2017

Revision as of 02:26, 6 January 2017

Line 1:

{{Regression bar}}

'''Nonparametric regression''' is a category of [[regression analysis]] in which the regression function does not take a predetermined form but is estimated according to information derived from the data. Nonparametric regression requires larger sample sizes than regression based on parametric models because the data must supply the model structure as well as the model estimates. The approach can handle both continuous and discrete variables for the dependent variable (called 'regressand') and for the explanatory variables (called 'regressors'), as well as allow for time dependency and dynamics.<ref>[http://www.sciencedirect.com/science/article/pii/S0167947316302596 Park, B.U., L. Simar, and V. Zelenyuk (2017). "Nonparametric estimation of dynamic discrete choice models for time series data," Computational Statistics and Data Analysis 108, pages 97-120.]</ref> and for the explanatory variables (called 'regressors')<ref>[https://ideas.repec.org/a/eee/csdana/v100y2016icp424-444.html Li, D., Simar, L. and V. Zelenyuk (2016) "Generalized nonparametric smoothing with mixed discrete and continuous data" Computational Statistics and Data Analysis, pages 424-444.]</ref>

==Gaussian process regression or Kriging==