Communications for Statistical Applications and Methods

The Korean Statistical Society (한국통계학회)
권호 표지

Prediction and Classification Using Projection Pursuit Regression with Automatic Order Selection

  • Published : 2000.08.01
Download PDF
⟨ Previous Next ⟩

Abstract

We developed a macro for prediction and classification using profection pursuit regression based on Friedman (1984b) and Hwang, et al. (1994). In the macro, the order of the Hermite functions can be selected automatically. In projection pursuit regression, we compare several smoothing methods such as super smoothing, smoothing with the Hermite functions. Also, classification methods applied to German credit data are compared.

Keywords

References

  1. Classification and regression trees Breiman, L.;Friedman, J.H.;Olshen, R.A.;Stone, C.J.
  2. Approximation Theory, Ⅵ Degree of nonlinear approximation DeVore, R.A.;C.K. Chui(ed.);L.L. Schumaker(ed.);D.J. Ward(ed.)
  3. The Annals of Statistics v.17 Projection-based approximation and a duality with kernel methods Donoho, D.L.;Johnstone, I.M.
  4. Laboratory for Computational Statistics Technical Report No.5 A variable span smoother Friedman, J.H.
  5. Laboratory for Computational Statistics Technical Report No.1 SMART user's guide Friedman, J.H.
  6. Journal of American Statistics Association v.76 Projection pursuit regression Friedman, J.H.;Stuetzle, W.
  7. IEEE Transaction on Neural Networks v.5 Regression modeling in back-propagation and projection pursuit learning Hwang, J.N.;Lay, S.R.;Maechler, M.;Martin, D.;Schimert, J.
  8. Learning From Data Cherkassky, Vladimir S.;Mulier, Filip M.
  9. NIPS4 Some approximation properties of projection pursuit learning networks Zhao, Y.;Atkeson, G.G.