gaum_bor_ln {noisemodel}

Gaussian-mixture borderline label noise

Description

Introduction of Gaussian-mixture borderline label noise into a classification dataset.

Usage

## Default S3 method:
gaum_bor_ln(
  x,
  y,
  level,
  mean = c(0, 2),
  sd = c(sqrt(0.5), sqrt(0.5)),
  w = c(0.5, 0.5),
  k = 1,
  sortid = TRUE,
  ...
)

## S3 method for class 'formula'
gaum_bor_ln(formula, data, ...)

Arguments

x

a data frame of input attributes.

y

a factor vector with the output class of each sample.

level

a double in [0,1] with the noise level to be introduced.

mean

a double vector with the mean for each Gaussian distribution (default: c(0,2)).

sd

a double vector with the standard deviation for each Gaussian distribution (default: c(sqrt(0.5),sqrt(0.5))).

w

a double vector with the weight for each Gaussian distribution (default: c(0.5,0.5)).

k

an integer with the number of nearest neighbors to be used (default: 1).

sortid

a logical indicating if the indices must be sorted at the output (default: TRUE).

...

other options to pass to the function.

formula

a formula with the output class and, at least, one input attribute.

data

a data frame in which to interpret the variables in the formula.

Details

Gaussian-mixture borderline label noise uses an SVM to induce the decision border in the dataset. For each sample, its distance to the decision border is computed. Then, a Gaussian mixture distribution with parameters (mean, sd) and weights w is used to compute the value for the probability density function associated to each distance. Finally, (level·100)% of the samples in the dataset are randomly selected to be mislabeled according to their values of the probability density function. For each noisy sample, the majority class among its k-nearest neighbors of a different class is chosen as the new label.

Value

An object of class ndmodel with elements:

xnoise

a data frame with the noisy input attributes.

ynoise

a factor vector with the noisy output class.

numnoise

an integer vector with the amount of noisy samples per class.

idnoise

an integer vector list with the indices of noisy samples.

numclean

an integer vector with the amount of clean samples per class.

idclean

an integer vector list with the indices of clean samples.

distr

an integer vector with the samples per class in the original data.

model

the full name of the noise introduction model used.

param

a list of the argument values.

call

the function call.

Note

Noise model adapted from the papers in References, considering SVM with linear kernel as classifier, a mislabeling process using the neighborhood of noisy samples and a noise level to control the number of errors in the data.

References

J. Bootkrajang and J. Chaijaruwanich. Towards instance-dependent label noise-tolerant classification: a probabilistic approach. Pattern Analysis and Applications, 23(1):95-111, 2020. doi:10.1007/s10044-018-0750-z.

See Also

gau_bor_ln, sigb_uni_ln, print.ndmodel, summary.ndmodel, plot.ndmodel

Examples

# load the dataset
data(iris2D)

# usage of the default method
set.seed(9)
outdef <- gaum_bor_ln(x = iris2D[,-ncol(iris2D)], y = iris2D[,ncol(iris2D)], level = 0.1)

# show results
summary(outdef, showid = TRUE)
plot(outdef)

# usage of the method for class formula
set.seed(9)
outfrm <- gaum_bor_ln(formula = Species ~ ., data = iris2D, level = 0.1)

# check the match of noisy indices
identical(outdef$idnoise, outfrm$idnoise)