discriminant.qda: Discriminant analysis under various assumed covariance matrix...

In tpepler/cpc: Common principal component (CPC) analysis and applications

Discriminant analysis under various assumed covariance matrix structures

Description

Performs discriminant analysis for two groups, under the assumption of equal, CPC or unrelated population covariance matrices.

Usage

discriminant.qda(origdata, group, method = c("unbiased", "pooled", "cpc", "fullcpccrossvalid"), B = NULL, standardize = FALSE)

Arguments

origdata	Matrix containing the sample data for two groups.
group	Vector (with values 1 and 2) indicating group membership for the rows in origdata.
method	Options: unbiased (default), cpc, fullcpccrossvalid, pooled. Each method estimates the covariance matrices of the two groups in a different way (under different assumptions). See Pepler (2014) for details.
B	Modal matrix, if available.
standardize	Logical, indicating whether the covariance matrices of the two groups should be calculated on standardised data before performing the discriminant analysis (default = FALSE). Discriminant analysis on the standardised covariance matrices is equivalent to discriminant analysis on the correlation matrices of the two groups.

Value

Returns a list with the following components:

inputdata	The observations in origdata, together with an indication of the group membership and the predicted group membership.
misclassrate	The misclasssification error rate.

Author(s)

Theo Pepler

References

Pepler, P.T. (2014). The identification and application of common principal components. PhD dissertation in the Department of Statistics and Actuarial Science, Stellenbosch University.

Examples

# Discriminant analysis to distinguish between the versicolor 
# and virginica groups, under the CPC assumption.
data(iris)
versicolor <- iris[51:100, 1:4]
virginica <- iris[101:150, 1:4]
S <- array(NA, dim = c(4, 4, 2))
S[, , 1] <- cov(versicolor)
S[, , 2] <- cov(virginica)
nvec <- c(nrow(versicolor), nrow(virginica))
B <- cpc::FG(covmats = S, nvec = nvec)$B
discriminant.qda(origdata = rbind(versicolor, virginica), group = c(rep(1, times = nrow(versicolor)), rep(2, times = nrow(virginica))), method = "cpc", B = B)

tpepler/cpc documentation built on July 7, 2022, 2:13 a.m.