CreateAlgorithm

Create algorithm

Syntax

NewAlg = CreateAlgorithm(Alg,AlgorithmName)

Description

NewAlg = CreateAlgorithm(Alg,AlgorithmName) creates an algorithm of the specified type.

Examples

Change Fit Algorithm

Change the fit algorithm for a model.

model = mbcmodel.CreateModel('Polynomial', 2);
minpress = model.FitAlgorithm.CreateAlgorithm('Minimize PRESS');
model.FitAlgorithm = minpress;

Display Algorithm Properties

AlgorithmName determines what properties you can set. Display the properties for an algorithm.

model.FitAlgorithm.properties

Algorithm: Minimize PRESS
Alternatives: 'Least Squares','Forward Selection','Backward 
Selection','Prune'
     MaxIter: Maximum Iterations (int: [1,1000])

Input Arguments

Alg — Algorithm

mbcmodel.fitalgorithm object

Algorithm, specified as a mbcmodel.fitalgorithm object.

AlgorithmName — Algorithm name

character vector

Algorithm name, specified as a character vector. AlgorithmName must be in the list of alternative algorithms given by getAlternativeNames.

Output Arguments

NewAlg — New algorithm

matrix

New algorithm, specified as a character vector.

Algorithms

AlgorithmName determines what properties you can set.

Linear Model Algorithms

Polynomials, hybrid splines, and the StepAlgorithm for RBF algorithms use these linear models.

Algorithm	Alternatives
Least Squares	'Minimize PRESS','Forward Selection','Backward Selection','Prune'
Minimize PRESS	'Least Squares','Forward Selection','Backward Selection','Prune'
Forward selection	'Least Squares','Minimize PRESS','Backward Selection','Prune'
Backward selection	'Least Squares','Minimize PRESS','Forward Selection','Prune'
Prune	'Least Squares','Minimize PRESS','Forward Selection','Backward Selection'

RBF Algorithms

For information about any of the RBF and Hybrid RBF algorithm properties, see Radial Basis Functions for Model Building and Fitting Routines.

Algorithm		Alternatives
Width Selection	TrialWidths	'Least Squares','Forward Selection','Backward Selection','Prune'
	WidPerDim	'TrialWidths','Tree Regression'
	Tree Regression	'TrialWidths','WidPerDim'
Lambda Selection	IterateRidge	'IterateRols','StepItRols'
	IterateRols	'IterateRidge','StepItRols'
	StepItRols	'IterateRidge','IterateRols'
Center Selection	Rols	'RedErr','WiggleCenters','CenterExchange'
	RedErr	'Rols','WiggleCenters','CenterExchange'
	WiggleCenters	'Rols','RedErr','CenterExchange'
	CenterExchange	'Rols','RedErr','WiggleCenters'
Tree Regression	Trial Alpha	'Specify Alpha'
	Specify Alpha	'Trial Alpha'
	Tree-based Center Selection	'Generic Center Selection'
	Generic Center Selection	'Tree-based Center Selection'

Hybrid RBF

Algorithm		Alternatives
Width Selection	TrialWidths	'WidPerDim','Tree Regression'
Nested Fit	Twostep	'Interlace'
	Interlace	'Twostep'

Boundary Model Fit

Algorithm		Alternatives
Convex Hull		'WidPerDim','Tree Regression'
Ellipsoid	Constraint Fitting	'Interlace'
	Boundary Points	'Twostep'
Star-shaped	Constraint Fitting
Star-Shaped – Special Points	Star-shaped Points	'Mean', 'Median', 'Mid Range', 'Min Ellipse', 'User Defined'
Star-Shaped – Boundary Points	Boundary Only (no further options)
Star-Shaped – Constraint Fit	Star-shaped RBF Fit
	Interpolation

Version History

Introduced in R2007a