NMFStrategy-class: Virtual Interface for NMF Algorithms
In renozao/NMF: Algorithms and Framework for Nonnegative Matrix Factorization (NMF)

Description Usage Arguments Functions Slots Methods (by generic) Examples

This class partially implements the generic interface defined for general algorithms defined in the NMF package (see algorithmic-NMF).

## S4 method for signature 'NMFStrategy'
show(object)

## S4 method for signature 'NMFStrategy,mMatrix,NMFfit'
run(object, y, x, ...)

## S4 method for signature 'NMFStrategy,mMatrix,NMF'
run(object, y, x, ...)

## S4 method for signature 'NMFStrategy'
objective(object)

## S4 replacement method for signature 'NMFStrategy,character'
objective(object) <- value

## S4 replacement method for signature 'NMFStrategy,'function''
objective(object) <- value

## S4 method for signature 'NMFStrategy'
modelname(object)

is.mixed(object)

## S3 method for class 'NMFStrategy'
nmfFormals(x, ...)

`y`	the target data that we want to approximate, passed as a matrix(-like) object.
`x`	an NMFfit object used as a starting point/seed by the algorithm.

is.mixed: is.mixed tells if an NMF algorithm works on mixed-sign data.

objective: the objective function associated with the algorithm (Frobenius, Kullback-Leibler, etc...). It is either an access key of a registered objective function or a function definition. In the latter case, the given function must have the following signature (x="NMF", y="matrix") and return a nonnegative real value.
model: a character string giving either the (sub)class name of the NMF-class instance used and returned by the strategy, or a function name.
mixed: a logical that indicates if the algorithm works on mixed-sign data.

canFit:

canFit(x = NMFStrategy,y = character): Tells if an NMF algorithm can fit a given class of NMF models
canFit(x = NMFStrategy,y = NMF): Tells if an NMF algorithm can fit the same class of models as y

deviance:

deviance(object = NMFStrategy): Computes the value of the objective function between the estimate x and the target y, using the objective function associated with the given NMFStrategy object.

modelname:

modelname(object = NMFStrategy): Returns the model(s) that an NMF algorithm can fit.

nmfFormals:

nmfFormals(NMFStrategy): Returns the extra arguments that can be passed to an algorithm encapsulated in an NMFStrategy object.

Arguments that have default values defined by the strategy are set accordingly.

NMFStrategy:

NMFStrategy(name = character,method = NMFStrategy): Creates an NMFStrategy object based on a template object (Constructor-Copy).
NMFStrategy(name = NMFStrategy,method = missing): Creates an NMFStrategy based on a template object (Constructor-Copy), in particular it uses the same name.
NMFStrategy(name = NULL,method = NMFStrategy): Creates an NMFStrategy based on a template object (Constructor-Copy) but using a randomly generated name.

objective:

objective(object = NMFStrategy): Gets the objective function associated with an NMF algorithm.

It is used in deviance to compute the objective value for an NMF model with respect to a given target matrix.

objective<-:

objective<-(object = NMFStrategy,value = character): Sets the objective function associated with an NMF algorithm, with a character string that must be a registered objective function.
objective<-(object = NMFStrategy,value = function): Sets the objective function associated with an NMF algorithm, with a function that computes the approximation error between an NMF model and a target matrix.

run:

run(object = NMFStrategy,y = mMatrix,x = NMFfit): Pure virtual method defined for all NMF algorithms to ensure that a method run is defined by sub-classes of NMFStrategy.

It throws an error if called directly.
run(object = NMFStrategy,y = mMatrix,x = NMF): Method to run an NMF algorithm directly starting from a given NMF model.

# get the type of model(s) associated with an NMF algorithm
modelname( nmfAlgorithm('brunet') )
modelname( nmfAlgorithm('nsNMF') )
modelname( nmfAlgorithm('offset') )