summarize {LogitBoost}R Documentation

Summarizes the output of crossval() and logitboost() by printing and plotting

Description

Prints and plots error-rates for optimal, fixed and (optionally) estimated stopping times when predicting a test set with logitboost(), or when running v-fold cross validation via crossval().

Usage

summarize(boost.out, resp, mout=100, grafik=T)

Arguments

boost.out A list, obtained as output of either logitboost() or crossval()
resp A numerical vector, containing the true response labels of the K classes as consecutive integers from 0 to (K-1).
mout An integer, giving the number of iterations of the boosting procedure, for which the error rate should be printed. The default value mout=100 is usually a good choice for gene expression data, which can well be inspected visually by the boosting error curve.
grafik Logical flag, indicates whether the boosting error curve should be plotted or not. The default is TRUE.

Value

Prints and plots the error-rates from the LogitBoost procedure.

Author(s)

Marcel Dettling

References

See "Boosting for Tumor Classification with Gene Expression Data", Dettling and Buhlmann (2002), available on the web page http://stat.ethz.ch/~dettling/boosting.html

Examples

data(leukemia)

## An example without stopping parameter estimation
fit <- crossval(leukemia.x,leukemia.y,v=5,mfinal=100,presel=75,verbose=TRUE)
summarize(fit, leukemia.y, grafik=FALSE)
summarize(fit, leukemia.y, mout=57)

[Package LogitBoost version 1.1-9 Index]