mvtboost-package: Tree Boosting for Multivariate Outcomes

In patr1ckm/mvtboost: Tree Boosting for Multivariate Outcomes

Description

Fits 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.

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 package vignettes:

vignette("mvtboost_wellbeing")

Author(s)

Patrick Miller [aut, cre], Daniel B. McArtor [aut]

Maintainer: Patrick Miller <[e-mail]>

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

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)

patr1ckm/mvtboost documentation built on May 24, 2019, 8:21 p.m.