Processing math: 0%

HYPbdp

HYPbdp finds constant c which is associated to the requested breakdown point for hyp. tan. estimator

expand all in page

Syntax

Description

example

c =HYPbdp(bdp, v) HYPbdp with all default options.

example

c =HYPbdp(bdp, v, k) Find value of c, A, B, for fixed break down point.

example

c =HYPbdp(bdp, v, k, traceiter) Find value of c, A, B, for fixed break down point.

example

[c, A] =HYPbdp(___) Efficiency monitoring.

example

[c, A, B] =HYPbdp(___) Analysis of efficiency as a function of k =sup CVC.

example

[c, A, B, d] =HYPbdp(___) Example of use of option traceiter.

Examples

expand all

  • HYPbdp with all default options.

    • Find required value of c when k=4.5 and bdp=0.5. 

    c=HYPbdp(0.5,1);

  • Find value of c, A, B, for fixed break down point.

    • Find required values when k=4.5 and bdp=0.5. 

    [c,A,B,d]=HYPbdp(0.5,1);
    % In this case
    % c = 2.010311082005501
    % A = 0.008931591866092
    % B = 0.051928487236632
    % d=  0.132017481327058

  • Find value of c, A, B, for fixed break down point.

    • Find required values when k=4.5 and bdp=0.5. 

    ktuning=4.2;
    [c,A,B,d]=HYPbdp(0.5,1,ktuning);
    % In this case
    % c =  2.093345907330513
    % A = 0.002411382994023
    % B = 0.026962054035832
    % d=  0.066725444793702

  • Efficiency monitoring.

    • Analysis of efficiency and of paramters A, B abd k as function of bdp for a given value of sup CVC=4 

    seqi=0.1:0.1:0.5;
    eff=[seqi' zeros(length(seqi),4)];
    iter=0;
    k=4;
    for i=seqi
    [c,A,B,d] = HYPbdp(i,1,k);
    iter=iter+1;
    eff(iter,2:5)=[B^2/A A B d];
    end
    subplot(2,2,1)
    plot(eff(:,1),eff(:,2))
    title('efficiency')
    subplot(2,2,2)
    plot(eff(:,1),eff(:,3))
    title('A')
    subplot(2,2,3)
    plot(eff(:,1),eff(:,4))
    title('B')
    subplot(2,2,4)
    plot(eff(:,1),eff(:,5))
    title('d')

  • Analysis of efficiency as a function of k =sup CVC.

    • ktun=[2; 3; 4; 5];
    eff=zeros(length(ktun),1);
    for i=1:length(ktun)
    [~,A1,B1]=HYPbdp(0.3,1,ktun(i));
    eff(i)=B1^2/A1;
    end
    % Efficiency increases as sup CVC increases.
    table(ktun,eff)

  • Example of use of option traceiter.

    • traceiter=1
    ktuning=4.5;
    [c,A,B,d]=HYPbdp(0.4,1,ktuning,traceiter);

    Related Examples

    expand all

  • Another example of use of option traceiter.

    • traceiter=-1 
    % traceiter=-1 implies that no messsage is shown on the screen about
    % effective tolerance in the iterative procedure.
    ktuning=4.5;
    [c,A,B,d]=HYPbdp(0.4,1,ktuning,traceiter);

    Input Arguments

    expand all

    bdp — requested breakdown point. Scalar.

    Scalar defining breakdown point (i.e a number between 0 and 0.5)

    Data Types: single|double

    v — number of response variables. Scalar.

    (e.g. in regression p=1)

    Data Types: single|double|int32|int64

    Optional Arguments

    k — supremum of the change of variance curve. Scalar.

    .

    Default value is k=4.5.

    Example: 5

    Data Types: double

    traceiter — Level of display. Scalar.

    If traceiter = 1 it is possible to monitor how the value of the objective function E(rho)/\rho(\infty) gets closer to the target (bdp) during the iterations If traceiter =0 just the message about effective convergence (if the number of iteration exceeds 80) is shown on the screen.

    If traceiter =-1 no message is shown on the screen.

    Example: 'traceiter',0

    Data Types: single|double|int32|int64

    Output Arguments

    expand all

    c —parameter c of hyperbolic tangent estimator. Scalar

    For more details see the methodological details inside "More About" below

    A —parameter A of hyperbolic tangent estimator. Scalar

    For more details see the methodological details inside "More About" below

    B —parameter B of hyperbolic tangent estimator. Scalar

    For more details see the methodological details inside "More About" below

    d —parameter d of hyperbolic tangent estimator. Scalar

    For more details see the methodological details inside "More About" below

    More About

    expand all

    Additional Details

    HYPpsi(u) = \left\{ \begin{array}{cc} u & |u| \leq d \\ \sqrt{A (k - 1)} \tanh \left( \sqrt{(k - 1) B^2/A} (c -|u|)/2 \right) sign(u) & d \leq |u| < c, \\ 0 & |u| \geq c. \end{array} \right. It is necessary to have 0 < A < B < 2 normcdf(c)-1- 2 c \times normpdf(c) <1

    References

    Hampel, F.R., Rousseeuw, P.J. and Ronchetti E. (1981), The Change-of-Variance Curve and Optimal Redescending M-Estimators, "Journal of the American Statistical Association", Vol. 76, pp. 643-648. [HRR]

    See Also

    | |

    This page has been automatically generated by our routine publishFS