Learn R Programming

mvtboost (version 0.5.0)

mvtboost-package: \Sexpr[results=rd,stage=build]{tools:::Rd_package_title("#1")}mvtboostTree Boosting for Multivariate Outcomes

Description

\Sexpr[results=rd,stage=build]{tools:::Rd_package_description("#1")}mvtboostFits a multivariate model of decision trees for multiple, continuous outcome variables. A model for each outcome variable is fit separately, selecting predictors that explain covariance in the outcomes. Built on top of 'gbm', which fits an ensemble of decision trees to univariate outcomes.

Arguments

Details

The most important function is mvtb, which fits the multivariate tree boosting model. See ?mvtb for details. The fitted model objects have summary, print, plot and predict methods. Additionally, mvtb.ri to computes the relative influence of each predictor, and mvtb.covex computes an estimate of the covariance explained in pairs of outcomes by predictors. These tables can be displayed as heatmaps using mvtb.heat. Examples for fitting, tuning and interpreting the models are available in the help pages and two package vignettes:

vignette("mvtboost_vignette") vignette("mvtboost_wellbeing")

References

Miller P.J., Lubke G.H, McArtor D.B., Bergeman C.S. (2015) Finding structure in data: A data mining alternative to multivariate multiple regression. Psychological Methods.

See Also

gbm

Examples

Run this code
data(wellbeing)
Y <- wellbeing[,21:26]
X <- wellbeing[,1:20]
Ys <- scale(Y)
cont.id <- unlist(lapply(X,is.numeric))
Xs <- scale(X[,cont.id])

res <- mvtb(Y=Ys,X=Xs)

summary(res)
plot(res,predictor.no = 8)
predict(res,newdata=Xs)

covex <- mvtb.covex(res, Y=Ys, X=Xs)
mvtb.cluster(covex)
par(mar=c(4,7,1,1))
mvtb.heat(covex,cexRow=.8)
par(mar=c(5,5,1,1))
mvtb.heat(t(mvtb.ri(res)),cexRow=.8,cexCol=1,dec=0)

Run the code above in your browser using DataLab