rlga

rlga performs robust linear grouping analysis

Syntax

• out=rlga(X,k,alpha)example
• out=rlga(X,k,alpha,Name,Value)example

Description

 out =rlga(X, k, alpha) rlga with all default options.

 out =rlga(X, k, alpha, Name, Value) rlga with niter = 500 and biter = 1000.

Examples

expand all

rlga with all default options.

rng(123); this leads to ROSS = 5.1298;

X=load('X.txt');
out=rlga(X,3,0.05);

rlga with niter = 500 and biter = 1000.

X=load('X.txt');
out=rlga(X,4,0.05,'niter',500,'biter',1000);

Related Examples

expand all

Generate mixture of regression using MixSimReg, with an average overlapping at centroids = 0.

01. Use all default options.

rng(372,'twister');
p=3;
k=2;
Q=MixSimreg(k,p,'BarOmega',0.001);
n=500;
[y,X,id]=simdatasetreg(n,Q.Pi,Q.Beta,Q.S,Q.Xdistrib);
% run rlga
out=rlga([y,X(:,2:end)],k,0.01);
RLGA Algorithm k =2 biter =8 niter =10
Finished.


Input Arguments

X — input data matrix. Matrix.

Input data as matrix of size n-by-p

Data Types: single| double

k — number of clusters. Scalar.

Scalar which specifies the number of clusters.

Data Types: single| double

alpha — a numeric value between 0 and 0.5. Scalar.

For the robust estimate of RLGA, specifying the percentage of points to be trimmed.

alpha must be a number in the interval [0 0.5].

Data Types: single| double

Name-Value Pair Arguments

Specify optional comma-separated pairs of Name,Value arguments. Name is the argument name and Value is the corresponding value. Name must appear inside single quotes (' '). You can specify several name and value pair arguments in any order as  Name1,Value1,...,NameN,ValueN.

Example:  'biter',1 , 'niter',1 , 'showall','true' , 'stand','true' , 'silent','true' , 'plots',1 

biter —number of different starting hyperplanes.scalar.

An integer for the number of different starting hyperplanes to try.

Example:  'biter',1 

Data Types: double

niter —Number of iterations.scalar.

An integer for the number of iterations to attempt for convergence.

Example:  'niter',1 

Data Types: double

showall —Level of display.logical.

If true then display all the outcomes, not just the best one.

Example:  'showall','true' 

Data Types: char

stand —Data standardization.boolean.

If true the X matrix is standardized using the standard deviation before fitting. Logical.

Example:  'stand','true' 

Data Types: char

silent —Text output.logical.

If true, produces no text output during processing. The default value is false.

Example:  'silent','true' 

Data Types: char

plots —plot on the screen.scalar.

If plots=1 a plot is showed on the screen with the final allocation (and if size(X,2)==2 with the lines associated to the groups).

Example:  'plots',1 

Data Types: double

Output Arguments

out — description Structure

Structure which contains the following fields

Value Description
cluster

vector containing the cluster memberships.

ROSS

the Residual Orthogonal Sum of Squares for the solution.

converged

logical. True if at least one solution has converged.

nconverg

the number of converged solutions (out of biter starts).

hpcoeff

best hyerplane

X

the (scaled if selected) dataset.

scaled

logical. Is the data set scaled?

k

the number of clusters to be found.

alpha

the trimming percentage.

biter

the biter setting used.

niter

the niter setting used.

class

'rlga'.

References

Garcia-Escudero, L.A., Gordaliza, A., San Martin, R., Van Aelst, S. and Zamar, R. (2009), Robust linear clustering, "Journal of the Royal Statistical Society: Series B", Vol. 71, pp. 301-318.

[doi: 10.1111/j.1467-9868.2008.00682.x]