mbcmodel.localmodel

Properties and methods for local model objects

Description

Use these properties and object functions to create and examine local model objects.

Creation

Syntax

LMObj = mbcmodel.CreateModel.(LocalModelType,Inputs)

Description

LMObj = mbcmodel.CreateModel.(LocalModelType,Inputs) creates an mbcmodel.localmodel object of local model type LocalModelType and inputs Inputs.

For a list of model types, see CreateModel.

Properties

Type — Type of model objects

vector

This property is read-only.

Type of model objects to be returned to mbcmodel.project, specified as a vector. model.Type returns the model type.

Note

Spaces and case in model Type are ignored.

The model type must be one in this table.

Type	Model Object
Polynomial	mbcmodel.linearmodel
Hybrid Spline	mbcmodel.linearmodel
RBF	mbcmodel.linearmodel
Hybrid RBF	mbcmodel.linearmodel
Polynomial-RBF	mbcmodel.linearmodel
Hybrid Spline-RBF	mbcmodel.linearmodel
Gaussian Process	mbcmodel.model
Free Knot Spline	mbcmodel.model
Transient	mbcmodel.model
User-Defined	mbcmodel.model
Neural Network	mbcmodel.model
Interpolating RBF	mbcmodel.model
Local Polynomial Spline	mbcmodel.localmodel
Local Polynomial with Datum	mbcmodel.localmodel
Local Polynomial	mbcmodel.localmodel
Local Hybrid Spline	mbcmodel.localmodel
Local Truncated Power Series	mbcmodel.localmodel
Local Free Knot Spline	mbcmodel.localmodel
Local Multiple Models	mbcmodel.localmodel
Local Growth	mbcmodel.localmodel
Local User-Defined	mbcmodel.localmodel
Local Transient	mbcmodel.localmodel
Local Average Fit	mbcmodel.localmodel

You can get a list of types by using getAlternativeTypes. Use this syntax.

Mlist = getAlternativeTypes(M)

In this syntax, M is an mbcmodel.model object.

Data Types: char | string

Status — Status of model fit

Not Fitted | Fitted | Best

This property is read-only.

Status of model fit, specified as Not Fitted, Fitted, or Best.

Data Types: char | string

NumInputs — Number of inputs to model

real positive scalar

This property is read-only.

Number of inputs to model, specified as a real positive scalar.

Data Types: double | single

InputData — Input training data

matrix

This property is read-only.

Input training data, specified as a matrix. InputData is specified when calling fit.

Data Types: double

OutputData — Output or response data

matrix

This property is read-only.

Output or response data, specified as a matrix. OutputData is specified when calling fit.

Data Types: double

Properties — Local model properties

mbcmodel.modelproperties object

Local model properties, specified as an mbcmodel.modelproperties object. Every local model object has an mbcmodel.modelproperties object. In this object, each local model type has specific properties, as described in these tables.

Local Polynomial Properties

Property	Description
Order	Polynomial order (vector int: {[0,Inf],2})
InteractionOrder	Maximum order of interaction terms (int: [0,Inf])
TransformInputRange	Transform inputs (Boolean)
ParameterNames	List of parameter names (read-only)
StepwiseStatus	Stepwise status {'Always','Never','Step'} (cell)
Transform	Transform function (char) or empty ('')
CovarianceModel	Covariance model (enum: {'None','Power','Exponential','Mixed'})
CorrelationModel	Correlation model (enum: {'None','MA(1)','AR(1)','AR(2)'})

Local Hybrid Spline Properties

Property	Description
Order	Spline and polynomial order (vector int: {[0,3],2})
SplineVariable	Spline variable
SplineInteraction	Order of interaction between spline and polynomial (int: [0,3])
Knots: Position of knots (vector real)	ParameterNames: List of parameter names (read-only)
StepwiseStatus	Stepwise status {'Always','Never','Step'} (cell)
Transform	Transform function (char) or empty ('')
CovarianceModel	Covariance model (enum: {'None','Power','Exponential','Mixed'})
CorrelationModel	Correlation model (enum: {'None','MA(1)','AR(1)','AR(2)'})

Local Polynomial Spline Properties

Property	Description
HighOrder	Polynomial order above knot (int: [2,Inf])
LowOrder	Polynomial order below knot (int: [2,Inf])
Transform	Transform function (char) or empty ('')
CovarianceModel	Covariance model (enum: {'None','Power','Exponential','Mixed'})
CorrelationModel	Correlation model (enum: {'None','MA(1)','AR(1)','AR(2)'})
DatumType	Datum Type (enum: {'None','Maximum','Minimum','Linked'})

Local Polynomial With Datum Properties

Property	Description
Order	Polynomial order (int: [0,Inf])
Transform	Transform function (char) or empty ('')
CovarianceModel	Covariance model (enum: {'None','Power','Exponential','Mixed'})
CorrelationModel	Correlation model (enum: {'None','MA(1)','AR(1)','AR(2)'})
DatumType	Datum type (enum: {'None','Maximum','Minimum','Linked'})

Local Free Knot Spline Properties

Property	Description
Order	Spline order (int: [0,Inf])
NumKnots	Number of knots (int: 'Positive')
Transform	Transform function (char) or empty ('')
CovarianceModel	Covariance model (enum: {'None','Power','Exponential','Mixed'})
CorrelationModel	Correlation model (enum: {'None','MA(1)','AR(1)','AR(2)'})

Local Truncated Power Series Properties

Property	Description
Order	Polynomial order (int: 'Positive')
NumKnots	Number of knots (int: 'Positive')
Transform	Transform function (char) or empty ('')
CovarianceModel	Covariance model (enum: {'None','Power','Exponential','Mixed'})
CorrelationModel	Correlation model (enum: {'None','MA(1)','AR(1)','AR(2)'})

Local Growth Properties

Property	Description
Model	Growth model (enum: {'expgrowth','gomp','logistic','logistic4','mmf','richards','weibul'})
AlternativeModels	List of growth models (read-only)
Transform	Transform function (char) or empty ('')
TransformBothSides	Transform both sides (Boolean)
CovarianceModel	Covariance Model (enum: {'None','Power','Exponential','Mixed'})
CorrelationModel	Correlation Model (enum: {'None','MA(1)','AR(1)','AR(2)'})

Local User-Defined Properties

Property	Description
Model	Name of user-defined model (enum: {'exponential'})
AlternativeModels	List of registered user-defined models (read-only)
Transform	Transform function (char) or empty ('')
TransformBothSides	Transform both sides (Boolean)
CovarianceModel	Covariance model (enum: {'None','Power','Exponential','Mixed'})
CorrelationModel	Correlation model (enum: {'None','MA(1)','AR(1)','AR(2)'})

Local Transient Properties

Property	Description
Model	Name of transient model (enum: {'fuelPuddle'})
AlternativeModels	List of registered transient models (read-only)
Transform	Transform function (char) or empty ('')
TransformBothSides	Transform both sides (Boolean)
CovarianceModel	Covariance model (enum: {'None','Power','Exponential','Mixed'})
CorrelationModel	Correlation model (enum: {'None','MA(1)','AR(1)','AR(2)'})

Local Multiple Models Properties

Property	Description
ModelCandidates	List of candidate models (cell)
SelectionStatistic	Selection statistic for automatic model selection (char). See the list for input names and descriptions. The list of valid statistics is the summary statistics in common with all model candidates. For example, if an interpolating RBF is one of the candidates, only RMSE is available.
AutomaticInputRanges	Use data range as model input ranges (Boolean)
Transform	Transform function (char) or empty ('')

Model Type	List of SelectionStatistic Inputs
Polynomial, Hybrid Spline, RBF, Hybrid RBF	'PRESS RMSE','RMSE','GCV','Weighted PRESS','-2logL','AIC','AICc','BIC','R^2','R^2 adj','PRESS R^2','DW','Cp','cond(J)'
Neural Network	'RMSE','R^2','R^2 adj','-2logL','AIC','AICc','BIC'
Free Knot Spline	'PRESS RMSE','RMSE','GCV','Weighted PRESS','-2logL','AIC','AICc','BIC','R^2','R^2 adj','PRESS R^2','DW','Cp'
Interpolating RBF	'RMSE'

SelectionStatistic Input Argument	Description
'PRESS RMSE'	Predicted standard error	'sqrt(PRESS/N)'
'RMSE'	Root mean squared error	'sqrt(SSE/(N-p))'
'GCV'	Generalized cross-validation variance	'N*SSE/(N-p)^2'
'Weighted PRESS'	Weighted predicted standard error	'sqrt(PRESS/(N-p-1))'
'-2logL'	-2 * log likelihood	'N*log(SSE/N)'
'AIC'	Akaike information criteria	'-2logL + 2*(p+1)'
'AICc'	Small sample Akaike information criteria	'-2logL + 2(p+1)*N/(N-p)'
'BIC'	Bayesian information criteria	'-2logL + 2*log(N)*(p+1)'
'R^2'	R^2	'1 - SSE/SST'
'R^2 adj'	Adjusted R^2	'1 - SSE/SST*(N-1)/(N-p)'
'PRESS R^2'	PRESS R^2	'1 - PRESS/SST'
'DW'	Durbin-Watson statistic	'sum((e_i-e_{i+1})^2)/sum(e_i^2) '
'Cp'	Mallow's statistic	'SSE/(SSEmax/(N-pmax)) - N + 2*p'
'cond(J)'	Condition of regression matrix	'cond(J)'

Local Average Fit Properties

Property	Description
Model	[1x1 mbcmodel.linearmodel]
Transform	Transform function (char) or empty ('')

Data Types: char | string

FitAlgorithm — Fit algorithm for model

mbcmodel.FitAlgorithm object

Fit algorithm for model, specified as an mbcmodel.model.FitAlgorithm object.

FitAlgorithm is a property of mbcmodel.model and of boundary model objects mbcboundary.AbstractBoundary and all subclasses.

An mbcmodel.model.FitAlgorithm object is contained within the FitAlgorithm property of an mbcmodel.model object or mbcboundary object.

As an alternative to using CreateAlgorithm, you can assign the algorithm name directly to the algorithm.

B.FitAlgorithm.BoundaryPointOptions = 'Boundary Only';

m.FitAlgorithm = 'Minimize PRESS';

Case and spaces are ignored.

To get a fitalgorithm object, F, from a model, use this code.

M = mbcmodel.CreateModel('Polynomial', 4);
F = M.FitAlgorithm

F = 
Algorithm: Least Squares
Alternatives: 'Minimize PRESS','Forward Selection','Backward 
Selection','Prune'
1x1 struct array with no fields.

Inputs — Model input

mbcmodel.modelinput object

Model input, specified as an mbcmodel.modelinput object.

Name — Model object name

character vector

Model object name, specified as a character vector.

Data Types: char | string

Units — Model output unit

vector

Model output unit, specified as a vector.

Data Types: double | single

Response — Response object

object

This property is read-only.

Response object in mbcmodel.project object, specified as an object.

IsBeingEdited — Boolean indicating whether model is being edited

true or 1 | false or 0

This property is read-only.

Boolean indicating whether model is being edited, specified as either true (1) or false (0).

Example: 0

Data Types: logical

ResponseFeatures — Response features for local model object

mbcmodel.responsefeatures object

Response features for local model object, specified as an mbcmodel.responsefeatures object. Each local model type has specific response features.

Data Types: char | string

Object Functions

fit	Fit model or boundary model to new or existing data, and provide summary statistics
SummaryStatistics	Summary statistics for response

Version History

Introduced before R2006a