search for: blackboost

Dear all, I am puzzled by R's memory usage when calling the blackboost function from package mboost to estimate a Gradient boosting model on a simulated dataset with 20 correlated variables and 100,000 obs. The blackboost object created by the function is only 15.3Mb, but R's memory usage increases by about 3.9Gb during the estimation of the model and the memory i...

mboost contains a blackboost method to build tree-based boosting models. I tried to write my own "cost-sensitive" ada family. But obviously my understanding to implement ngradient, loss, and offset functions is not right. I would greatly appreciate if anyone can help me out, or show me how to write a cost-sensitive f...

...dels that would perform comparably. This, however, is not the case - mboost performs much better. This seems odd. This issue is meaningful since the output of this regression needs to be implemented in a production system, and mboost doesn't even expose the ensembles. # default params for blackboost are a gaussian loss, and maxdepth of 2 m.mboost = blackboost(Y ~ X1 + X2, data=tdata, weights=t.ipcw, control=boost_control(mstop=100)) m.gbm = gbm(Y ~ X1 + X2, data=tdata, weights=t.ipcw, distribution="gaussian", interaction.depth=2, bag.fraction=1, n.trees=2500) # compare IPCW weighted...

...to avoid name collisions between packages (eg. gam(), splsda(), roc(), LogitBoost()). Also search Bioconductor. (9) Allow the predict function to generate results from many different sub-models simultaneously. For example, pls() can return predictions across many values of ncomp. enet(), cubist(), blackboost() are other examples. corollary (9a): [OCD] ensure the same object type for predictions. There are occasions where predict() will return a vector or a matrix depending on the context. I would argue that this is not optimal. (10) Use a limited vocabulary for options. For example, some predict() fu...

...caret tries to avoid re-fitting models if it can get predictions from sub-models. For example, an object for a boosted tree with 500 trees can often be used to get predictions for any boosted tree with less than 500 trees. The affected models are: pls, plsda, earth, rpart, gbm, gamboost, glmboost, blackboost, ctree, pam, enet and lasso. The caretLSF package is a parallel processing version of caret. The other caret package, caretNWS, will be updated to work with the new version of caret shortly. Please email me at max dot kuhn at pfizer dot com with any questions or comments Max __________________...

...es are available. The project is now hosted on R-Forge. The homepage is http://caret.r-forge.r-project.org/ The package currently includes model tuning/resampling for the following models: lm, single trees (C4.5, rpart, ctree, logistic model trees), mars (via earth), boosted models (ada, gbm, blackboost, glmboost, gamboost, logitboost), bagged models (trees, earth, fda), randomforests (randomforest and cforest), rule-based models (Ripper and M5 prime), discriminant models (lda, fda, rda, ssda, slda), kernel methods (lssvm, ksvm, rvm, gausspr), nnet, nnet with initial pca step, multinom, pls, plsda...

...caret tries to avoid re-fitting models if it can get predictions from sub-models. For example, an object for a boosted tree with 500 trees can often be used to get predictions for any boosted tree with less than 500 trees. The affected models are: pls, plsda, earth, rpart, gbm, gamboost, glmboost, blackboost, ctree, pam, enet and lasso. The caretLSF package is a parallel processing version of caret. The other caret package, caretNWS, will be updated to work with the new version of caret shortly. Please email me at max dot kuhn at pfizer dot com with any questions or comments Max __________________...

...es are available. The project is now hosted on R-Forge. The homepage is http://caret.r-forge.r-project.org/ The package currently includes model tuning/resampling for the following models: lm, single trees (C4.5, rpart, ctree, logistic model trees), mars (via earth), boosted models (ada, gbm, blackboost, glmboost, gamboost, logitboost), bagged models (trees, earth, fda), randomforests (randomforest and cforest), rule-based models (Ripper and M5 prime), discriminant models (lda, fda, rda, ssda, slda), kernel methods (lssvm, ksvm, rvm, gausspr), nnet, nnet with initial pca step, multinom, pls, plsda...

...abase. This is why I highly value the fact that mboost has > glmboost, which can return linearly interpretable coefficients. It is > also why I do not directly call upon R every time I want to query a > model. (As an aside, if I were to use R directly, I might consider the > gamboost or blackboost methods, which do not return scalar > coefficients that are readily extractable.) > > > > On Sun, Feb 7, 2010 at 6:31 PM, David Winsemius <dwinsemius at > comcast.net> wrote: > > > > On Feb 7, 2010, at 5:03 PM, Kyle Werner wrote: > > > >> I...