Processing math: 0%

Mlocation

Mlocation finds the M estimator of location in a univariate sample

expand all in page

Syntax

  • newloc=Mlocation(y, psifunc)example
  • newloc=Mlocation(y, psifunc, scaleest)example
  • newloc=Mlocation(y, psifunc, scaleest, initialmu)example
  • newloc=Mlocation(y, psifunc, scaleest, initialmu, tol)example
  • newloc=Mlocation(y, psifunc, scaleest, initialmu, tol, maxiter)example

Description

example

newloc =Mlocation(y, psifunc) Example of M estimate of location.

example

newloc =Mlocation(y, psifunc, scaleest) Example of use of option scaleest.

example

newloc =Mlocation(y, psifunc, scaleest, initialmu) Example of use of option initialmu.

example

newloc =Mlocation(y, psifunc, scaleest, initialmu, tol) Example of use of options tol and maxiter.

example

newloc =Mlocation(y, psifunc, scaleest, initialmu, tol, maxiter) Compare location estimate using two different link functions.

Examples

expand all

  • Example of M estimate of location.

    • M estimate of location using Tukey biweight rho function with a value of c associated to a breakdown point of 0.5. 

    psifunc=struct;
    psifunc.class='TB';
    bdp=0.5;
    c=TBbdp(bdp,1);
    psifunc.c1=c;
    n=100;
    shift=100;
    u=randn(n,1);
    u(1:10)=u(1:10)+shift;
    % Function Mlocation is just called with two input arguments and
    % therefore the fixed estimate of the scale which is used is the scaled
    % MAD
    loc=Mlocation(u,psifunc)
    disp('Non robust estimate of location')
    disp(mean(u))
    disp('Robust estimate of location')
    disp(loc)

  • Example of use of option scaleest.

    • M estimate of the location using Hampel rho function with a value of c associated to a breakdown point of 0.5 

    psifunc=struct;
    psifunc.class='HA'
    abc=[1.5 3.5 8];
    bdp=0.5;
    c=HAbdp(bdp,1,abc);
    psifunc.c1=[c abc];
    n=10000;
    shift=50;
    truescale=6;
    u=truescale*randn(n,1);
    u(1:10)=u(1:10)+shift;
    % Mlocation is called with three arguments. 
    % The third argument is the fixed value of the scale to use in the loop.
    locHA=Mlocation(u,psifunc,truescale+1)

  • Example of use of option initialmu.

    • Estimate of location using Hampel rho function. 

    % M estimate of the location using Hampel rho function with a
    % value of c associated to a breakdown point of 0.5
    psifunc=struct;
    psifunc.class='HA'
    abc=[1.5 3.5 8];
    bdp=0.5;
    c=HAbdp(bdp,1,abc);
    psifunc.c1=[c abc];
    n=10000;
    shift=50;
    truescale=3;
    u=truescale*randn(n,1);
    u(1:10)=u(1:10)+shift;
    % Mlocation is called with four arguments.
    % The fourth argument is the initial estimate of mu to use.
    % Use as initial estimate of mu a trimmed mean.
    initialmu=trimmean(u,60);
    locHA=Mlocation(u,psifunc,truescale,initialmu)

  • Example of use of options tol and maxiter.

    • M estimate of the scale using Tukey biweight rho function with a value of c associated to a breakdown point of 0.5. 

    psifunc=struct;
    psifunc.class='TB';
    bdp=0.5;
    c=TBbdp(bdp,1);
    % kc = E(rho) = sup(rho)*bdp
    kc=c^2/6*bdp;
    psifunc.c1=c;
    psifunc.kc1=kc;
    n=10000;
    shift=5;
    u=2*randn(n,1);
    u(1:10)=u(1:10)+shift;
    % Mlocation is called with 6 arguments.
    % The fifth and sixth arguments are respectively tol and maxiter
    loc=Mlocation(u,psifunc,3,median(u),1e-7,20)

  • Compare location estimate using two different link functions.

    • psifunc=struct;
    psifunc.class='HA'
    abc=[1.5 3.5 8];
    bdp=0.5;
    c=HAbdp(bdp,1,abc);
    psifunc.c1=[c, abc];
    n=1000;
    shift=100;
    u=3*randn(n,1);
    u(100:200)=u(100:200)+shift;
    sHA=Mlocation(u,psifunc)
    psifunc=struct;
    psifunc.class='TB';
    c=TBbdp(bdp,1);
    psifunc.c1=c;
    sTB=Mlocation(u,psifunc);
    sMLE=mean(u);
    cate=categorical({'Robust location (Hampel)' 'Robust location (TB)' 'Non robust location'});
    bar(cate,[sHA sTB sMLE])

    Input Arguments

    expand all

    y — Response variable. Vector.

    n-by-1 vector of observations

    Data Types: single | double

    psifunc — rho (psi) function. Structure.

    A structure specifying the class of rho (psi) function to use and the consistency factor.

    psifunc must contain the following fields:

    Value Description
    class

    string identyfing the rho (psi) function to use.

    Admissible values for class are 'bisquare' (TB), 'optimal', (OPT) 'hyperbolic' (HYP), 'hampel' (HA) 'power divergence' (PD)
    c1

    consistency factor (and other parameters) associated to required breakdown point or nominal efficiency.

    More precisely, psifunc.c1(1) contains consistency factor associated to required breakdown point or nominal efficiency psifunc.c1(2:end) may contain other parameters associated with the rho (psi) function.

    For example, if psifunc.class='hampel', c1(2:4) may contain parameters (a, b and c) of Hampel rho function.

    Example: Example - psifunc.class='TB';psifunc.c1=1.5476;

    Data Types: struct

    Optional Arguments

    scaleest — value of sigma. Scalar.

    The estimate of the scale to use in the iterative loop.

    If not defined, scaled MAD of vector y is used.

    Example: 0.34

    Data Types: double

    initialmu — starting value of the location estimate. Scalar.

    The initial estimate of location to use in the first iteration.

    If not defined, initialmu is set equal to median(y).

    Example: trimmean(y,30)

    Data Types: double

    tol — scalar. The tolerance for controlling convergence.

    If not defined, tol is fixed to 1e-7.

    Example: 1e-10

    Data Types: double

    maxiter — scalar. Maximum number of iterations to find the location estimate.

    If not defined, maxiter is fixed to 200.

    Example: 100

    Data Types: double

    Output Arguments

    expand all

    newloc —M-estimate of location. Scalar

    Robust M estimate of location.

    More About

    expand all

    Additional Details

    Remark: the M estimator of location must satisfy the following equation If \rho is differentiable, with \psi(u) = \rho'(\psi) = \text{d}\psi(u)/\text{d}u, \hat{\mu} is the solution of the equation (estimating equation) \sum_{i=1}^n \psi \left( \frac{y_i - \hat{\mu}}{\sigma} \right) = 0. This estimating equation shows the importance of \psi(y). The equation can be rewritten to provide an algorithm for estimation of \mu as \sum_{i=1}^n w_i(y_i - \hat{\mu}) = 0 \quad \text{where} \quad w_i = \psi(y_i - \hat{\mu})/(y_i - \hat{\mu}).

    This routine computes the value of \mu which satisfies the above equation. \sigma corresponds to optional input parameter scaleest.

    Note that the value of \sigma is kept fixed in each iteration.

    The iterative procedure starts with some \hat{\mu}_0, for example the sample median.

    In this routine \hat{\mu}_0 corresponds to optional input parameter initialmu.

    Given the estimate \hat{\mu}_k at stage k, compute

    w_{i,k} = w(y_i - \hat{\mu}_k), \quad i = 1,\ldots,n \hat{\mu}_{k+1} = \sum_{i=1}^n w_{i,k}y_i/\sum_{i=1}^n w_{i,k}.

    References

    Maronna, R.A., Martin D. and Yohai V.J. (2006), "Robust Statistics, Theory and Methods", Wiley, New York.

    See Also

    | |

    This page has been automatically generated by our routine publishFS