flashData: Generation of a Binary Classification Problem with a...

In locClassData: Collection of Artificial Classification Problems

Description

Generation of a binary classification problem with a lightning-shaped decision boundary.

Usage

  flashData(n, prior = rep(0.5, 2), lambda = rep(0.5, 2),
    sigma = diag(2))

  flashLabels(data, prior = rep(0.5, 2),
    lambda = rep(0.5, 2), sigma = diag(2))

  flashPosterior(data, prior = rep(0.5, 2),
    lambda = rep(0.5, 2), sigma = diag(2))

  flashBayesClass(data, prior = rep(0.5, 2),
    lambda = rep(0.5, 2), sigma = diag(2))

Arguments

n	Number of observations.
prior	Vector of class prior probabilities.
lambda	The conditional probabilities for the mixture components given the class. Either a vector (if the same number m of mixture components is desired for each class and the conditional probabilities for each class should be equal) or a list as long as the number of classes containing one vector of probabilities for every class. The length of the k-th element is the desired number of mixture components for the k-th class.
sigma	The covariance matrices of the mixture components. Either one single matrix that is used for each mixture component or a list as long as the number of classes. List elements can be matrices (in case that for all mixture components forming one class the same covariance matrix shall be used) or lists of matrices as long as the number of mixture components in the corresponding class.
data	A data.frame.

Value

flashData returns an object of class "locClass", a list with components:

x	(A matrix.) The explanatory variables.
y	(A factor.) The class labels.

flashLabels returns a factor of class labels.

flashPosterior returns a matrix of posterior probabilities.

flashBayesClass returns a factor of Bayes predictions.

locClassData documentation built on May 2, 2019, 5:26 p.m.