threeGaussians: Two-dimensional toy data consisting of four Gaussian...

In benmack/oneClass: One-class classification in the absence of test data

Usage

threeGaussians(n = 200, d = 2:4, n.tr = c(20, 100), seed = NULL)

Arguments

n	integer. the number of test samples
d	vector of three integers. the distance between the three negative gaussians and the positive gaussian
n.tr	vector of two integer, i.e. the number of positive and unlabeled training samples
seed	a seed point controlling the randomness.

Value

a list of the toy data set with two data frames (tr and te) both containing the columns y, x1, and x2. in the training set the column y in {0, 1} indicates if a sample belongs to the positive class (1) or if it is unlabeled (0), and in the test set te if it belongs to the positive class (1) or the negative class (0). x1 and x2 are the features or predictors.

Examples

## Not run: 
toy <- threeGaussians(seed=9)

### plot the test set
par(mfrow=c(1,2))
plot( toy$te[ , -1 ], pch = c( 4, 16 )[ toy$te$y + 1 ] )
legend('topright', legend=c('pos', 'neg'), pch=c(16, 4))

### plot the training set
plot( toy$tr[ , -1 ], pch = c( 4, 16 )[ toy$tr$y + 1 ] )
legend('topright', legend=c('pos', 'un'), pch=c(16, 4))

## End(Not run)

benmack/oneClass documentation built on Dec. 15, 2020, 7:38 p.m.