TFunHDDC (version 1.0.2)

predict.tfunHDDC: predict.tfunHDDC: Predicting Function for t-funHDDC Objects

Description

Provides the matrix of classification probabilities and the classification vector for inputted observations assuming the model provided by the tfunHDDC object.

Usage

# S3 method for tfunHDDC
predict(object, data=NULL, ...)

Value

t

Matrix of classification probabilities

class

Vector of maximum a posteriori classifications

Arguments

object

An object of class tfunHDDC

data

Data frame (univariate funtional data) or a list (multivariate functional data) of new observations on the same variables used in the fitting of the tfunHDDC object. If NULL, then the observations used in the fitting of the tfunHDDC object are inputted.

...

Arguments to be passed to other functions.

Author

Cristina Anton, Iain Smith

References

-Andrews JL, McNicholas PD (2012) Model-based clustering, classification, and discriminant analysis via mixtures of multivariate t-distributions: The teigen family. Stat Comput 22:1021-1029. <doi.org/10.1007/ s11222-011-9272-x>

-Andrews JL, Wickins JR, Boers NM, et al (2018) An R package for modelbased clustering and classification via the multivariate t distribution. Journal of Statistical Software 83(7):1-32

- Cristina Anton, Iain Smith (2024) Model-based clustering of functional data via mixtures of \(t\) distributions. Advances in Data Analysis and Classification 18(3): 563-595.

See Also

tfunHDDC

Examples

# \donttest{
set.seed(1027)
#simulataed univariate data

data = genModelFD(ncurves=300, nsplines=35, alpha=c(0.9,0.9,0.9),
                  eta=c(10, 7, 17))

plot(data$fd, col = data$groupd)

clm = data$groupd

model1=c("AkjBkQkDk", "AkjBQkDk", "AkBkQkDk", "ABkQkDk", "AkBQkDk", "ABQkDk")

####################classification example with predictions

training=c(1:50,101:150, 201:250)

test=c(51:100,151:200, 251:300)

known1=clm[training]

t4<-tfunHDDC(data$fd[training],K=3,threshold=0.2,init="kmeans",nb.rep=1,
             dfconstr="no", dfupdate="numeric", model=model1[1],known=known1, 
             itermax = 10)

if (!is.null(t4$class)) {
	table(clm[training], t4$class)

	p1<-predict.tfunHDDC(t4,data$fd[test] )

	if (!is.null(p1$class)) table(clm[test], p1$class)
}

###########################NOX data

data1=fitNOxBenchmark(15)

plotNOx(data1)

###example for prediction

training=c(1:50)

test=c(51:115)

known1=data1$groupd[training]

t1<-tfunHDDC(data1$fd[training],K=2,threshold=0.6,init="kmeans",nb.rep=10,
             dfconstr="no", model=c("AkjBkQkDk", "AkjBQkDk", "AkBkQkDk", 
             "ABkQkDk", "AkBQkDk", "ABQkDk"),known=known1) 

if (!is.null(t1$class)) {
	table(data1$groupd[training], t1$class)

	p1<-predict.tfunHDDC(t1,data1$fd[test] )

	if (!is.null(p1$class)) table(data1$groupd[test], p1$class)
}
# }