Rdimtools (version 1.0.0)

do.udp: Unsupervised Discriminant Projection

Description

Unsupervised Discriminant Projection (UDP) aims finding projection that balances local and global scatter. Even though the name contains the word Discriminant, this algorithm is unsupervised. The term there reflects its algorithmic tactic to discriminate distance points not in the neighborhood of each data point. It performs PCA as intermittent preprocessing for rank singularity issue. Authors clearly mentioned that it is inspired by Locality Preserving Projection, which minimizes the local scatter only.

Usage

do.udp(
  X,
  ndim = 2,
  type = c("proportion", 0.1),
  preprocess = c("center", "scale", "cscale", "decorrelate", "whiten")
)

Arguments

X

an \((n\times p)\) matrix or data frame whose rows are observations and columns represent independent variables.

ndim

an integer-valued target dimension.

type

a vector of neighborhood graph construction. Following types are supported; c("knn",k), c("enn",radius), and c("proportion",ratio). Default is c("proportion",0.1), connecting about 1/10 of nearest data points among all data points. See also aux.graphnbd for more details.

preprocess

an additional option for preprocessing the data. Default is "center". See also aux.preprocess for more details.

Value

a named list containing

Y

an \((n\times ndim)\) matrix whose rows are embedded observations.

trfinfo

a list containing information for out-of-sample prediction.

projection

a \((p\times ndim)\) whose columns are basis for projection.

interimdim

the number of PCA target dimension used in preprocessing.

References

yang_globally_2007Rdimtools

See Also

do.lpp

Examples

# NOT RUN {
## use iris data
data(iris)
X     = as.matrix(iris[,1:4])
label = as.integer(iris$Species)

## use different connectivity level
out1 <- do.udp(X, type=c("proportion",0.05))
out2 <- do.udp(X, type=c("proportion",0.10))
out3 <- do.udp(X, type=c("proportion",0.25))

## visualize
opar <- par(no.readonly=TRUE)
par(mfrow=c(1,3))
plot(out1$Y, col=label, main="connectivity 5%")
plot(out2$Y, col=label, main="connectivity 10%")
plot(out3$Y, col=label, main="connectivity 25%")
par(opar)
# }
# NOT RUN {
# }