ASpsix

ASpsix computes psi function (derivative of rho function) times x for Andrew's sine function

Syntax

Description

example

psix =ASpsix(u, c) Plot of psi function (derivative of rho function) times x for Andrews weight function.

Examples

expand all

  • Plot of psi function (derivative of rho function) times x for Andrews weight function.
  • x=-6:0.01:6;
    psixAS=ASpsix(x,2);
    plot(x,psixAS)
    xlabel('x','Interpreter','Latex')
    ylabel('$\psi (x)$','Interpreter','Latex')

    Input Arguments

    expand all

    u — scaled residuals or Mahalanobis distances. Vector.

    n x 1 vector containing residuals or Mahalanobis distances for the n units of the sample

    Data Types: single| double

    c — tuning parameters. Scalar.

    Scalar greater than 0 which controls the robustness/efficiency of the estimator (beta in regression or mu in the location case ...)

    Data Types: single| double

    More About

    expand all

    Additional Details

    Function ASpsix transforms vector u as follows \[ ASpsix(u)= \left\{ \begin{array}{cc} (u/c) sin(u/c) & |u/c| \leq \pi \\ 0 & |u/c|> \pi \\ \end{array} \right. \]

    Remark: Andrew's psi-function is almost linear around u = 0 in accordance with Winsor's principle that all distributions are normal in the middle.

    This means that \psi (u)/u is approximately constant over the linear region of \psi, so the points in that region tend to get equal weight.

    References

    Andrews, D.F., Bickel, P.J., Hampel, F.R., Huber, P.J., Rogers, W.H., and Tukey, J.W. (1972), "Robust Estimates of Location: Survey and Advances", Princeton Univ. Press, Princeton, NJ. [p. 203]

    Andrews, D. F. (1974). A Robust Method for Multiple Linear Regression, "Technometrics", V. 16, pp. 523-531, https://doi.org/10.1080/00401706.1974.10489233

    This page has been automatically generated by our routine publishFS