R Shiny app: Stochastic gradient boosting with gbm
I am working on a new R Shiny web application, which allows the user to perform stochastic gradient boosting using the R gbm package.
The app currently only uses simulated data, which come directly from the
gbm function help file. Currently, features are limited. Also, the app will throw an error if you try to perform cross-validation when gradient boosting, i.e., by setting the number of cross-validation folds > 1. This meta-parameter setting corresponds to the
cv.folds argument to
gbm. The help file example will run fine in a stand-alone R session, but not when built into a Shiny app. I speculate that this has something to do with the current version of the gbm package attempting to take advantage of the base R parallel package, but I have not been able to solve the problem. If anyone else can figure it out, please let me know!
This app is a very rough work in progress, but I wanted to share what I’ve got for now. I would like to continue enhancing this app with many additional features and graphics, not to mention switching over from the simulated dataset to something more real and relevant. Stay tuned for updates.