dataf.sim.1.CFF07 {ddalpha}

Model 1 from Cuevas et al. (2007)

Description

Model 1 from Cuevas et al. (2007)

Processes:
X(t) = m_0(t) + e(t), m_0(t) = 30*(1-t)*t^1.2
Y(t) = m_1(t) + e(t), m_1(t) = 30*(1-t)^1.2*t
e(t): Gaussian with mean = 0, cov(X(s), X(t)) = 0.2*exp(-abs(s - t)/0.3)
the processes are discretized at numDiscrets equally distant points on [0, 1]. The functions are smooth and differ in mean only, which makes the classification task rather simple.

Usage

dataf.sim.1.CFF07(numTrain = 100, numTest = 50, numDiscrets = 51, plot = FALSE)

Arguments

numTrain

number of objects in the training sample

numTest

number of objects in the test sample

numDiscrets

number of points for each object

plot

if TRUE the training sample is plotted

Format

A data strusture containing $learn and $test functional data. The functional data are given as data structures.

dataf

The functional data as a list of objects. Each object is characterized by two coordinates.

args

a numeric vector

vals

a numeric vector

labels

The classes of the objects: 0 for X(t), 1 for Y(t)

Source

Cuevas, A., Febrero, M. and Fraiman, R. (2007). Robust estimation and classification for functional data via projection-based depth notions. Computational Statistics 22 481-496.

See Also

dataf.* for other functional data sets

plot.functional for building plots of functional data

Examples

## load the dataset
dataf = dataf.sim.1.CFF07(numTrain = 100, numTest = 50, numDiscrets = 51)
learn = dataf$learn
test = dataf$test

## view the classes
unique(learn$labels)

## access the 5th point of the 2nd object
learn$dataf[[2]]$args[5]
learn$dataf[[2]]$vals[5]

## Not run: 
## plot the data
plot(learn)
plot(test)

## or
dataf = dataf.sim.1.CFF07(numTrain = 100, numTest = 50, numDiscrets = 51, plot = TRUE)

## End(Not run)