marginalRelevance: Marginal Relevance

In BKPC: Bayesian Kernel Projection Classifier

Description

Calculates Marginal Relevance of each feature (variable) useful for class (group) separation. The marginal relevance score is a ratio of the between-group to within-group sum of squares.

Usage

marginalRelevance(x, y)

Arguments

x	a data matrix.
y	a response vector. Should be a factor.

Value

An object of class "marginalRelevance" including:

score	Marginal relevance score of each feature.
rank	The ranking in order of highest marginal relevance for each feature.
orderedData	Data matrix with columns reordered by the marginal relevance of the features.
bestVars	Features ordered by the marginal relevance.

Author(s)

K. Domijan

References

Dudoit S., J. Fridlyand, T. P. Speed: Comparison of discrimination methods for the classification of tumors using gene expression data. Journal of the American Statistical Association, 2002, Volume 97 No 457, pp 77-87.

See Also

plot.marginalRelevance microarray

Examples

data(microarray)

profiles <- as.matrix(microarray[, -2309])
tumourType <-  microarray[, 2309]
 
 margRelv <- marginalRelevance(profiles, tumourType)

# plot 30 gene profiles with highest marginal relevance score
 plot(margRelv, type = "parallelcoord", n.feat = 50, col = tumourType )

 
## Not run: 
# another example: wine data from gclus
library(gclus)
data(wine)
dt <- as.matrix(wine[, -1])
colnames(dt) <- names(wine[, -1])

label <- as.factor(wine[, 1])


margRelv <- marginalRelevance(dt, label)

#variables in order of their MR score
colnames(dt)[ margRelv$bestVars]


cparcoord(dt, order = margRelv$bestVars, col = label)
cpairs(dt, order = margRelv$bestVars, col = label)


## End(Not run)

BKPC documentation built on May 1, 2019, 9:10 p.m.