seeg (version 1.0)

LDA: LDA Two groups

Description

Functions for linear discriminant analysis between two groups of observations for a set of variables.

Usage

lda2(X1,X2)

Arguments

X1

Group 1 data matrix

X2

Group 2 data matrix

Value

• m.n1.n2
  m.n1.n2
• G1
  sample mean of X1
• G2
  sample mean of X2
• S1
  covariance matrix of X1
• S2
  covariance matrix of X2
• Sp
  pooled variance matrix
• Spinv
  inverse of pooled variance matrix
• A
  matrix of coefficients
• Z1c.Z2c
  scores of centroids for group 1 and group 2
• D2.F.p.value
  Mahalanobis distance, F statistic, and p-value
• Z1s
  scores for individual observations of group 1
• Z2s
  scores for individual observations of group 2

Details

The centroid scores Z1c, Z2c and individual scores Z1s, Z2s have been centered on Z0, which is the midpoint or average of the centroids. Null hypothesis is that there are not significant differences between the two groups.

Using the individual scores and the centroid scores, the function also produces a plot, which displays the difference between the groups in two different manners. The one at the top panel uses markers and the two below use histograms. It indicates whther tere is overlap in observations along the score axis and potential differences between the centroids.

References

Acevedo M.F. 2013. "Data Analysis and Statistics for Geography, Environmental Science, and Engineering", CRC Press.

See Also

mda package lda

Examples

X1X2 <- read.csv("lab10/X1X2.csv",header=T)
X1 <- as.matrix(X1X2[,1:3]); X2 <- as.matrix(X1X2[,4:6])
X.lda2 <- lda2(X1,X2)