drmerlot/startml

R package for automatic hyper parameter tuning and ensembles with deep learning, gradient boosting machines, and random forests. Powered by h2o.

startml = start machine learning

Start doing machine learning with training optimization and ensembles of popular machine learning algorithms. Powered by the scalable implenentations of algoriths provided by h2o.

Why startml? Anyone who really wants to build machine learning models and apply them in a valuable way to real-word problems can with the help of open source technology.

What startml does is speed up the learning-through-doing process for new comers.

How this happens is by providing tools for automatically building many models, helping you put some miles on your CPU doing machine learning, and hopefully, giving you a little boost to some early success, inspiring more effort towards building a valuable knowledge and skill set.

It might be for you if: You want to do machine learning, you are excited to learn new things, you are focused on the fastest path to creating value, and you are (a little) impatient.

Install

startml is an R package currently hosted here on github. You can install it with devtools from this repository.

If you don't have devtools already, then in an R instance or Rstudio:

# First get devtools
install.packages("devtools")
library(devtools) 

# Then install startml: 
install_github("andrewsommerlot/startml")
library(startml)

startml's Issues

get_prediction function returns 1 member list. Needs to be subsetted again

Subset predictions within the get_prediction function so it doesnt return an object that needs to be subbsetted again.

Need file paths to not reference wd in demo function

Right now demo will fail if user not in project working directory

plot function does not work on mlblob of binomial models

select_models function fails when total number of models is less than 2

Error in base::colnames<-(x, value) : attempt to set 'colnames' on an object with less than two dimensions

5. stop("attempt to set 'colnames' on an object with less than two dimensions")
6. base::colnames<-(x, value)
7. colnames<-(*tmp*, value = 1L) at select-models.R#29
   ... startml

Hyperparameter search options need to be settable in all three algorithm grid search functions

return_dataframe option in startml function does not create dataframe

Need to add hierarchical hyperparameter arguments to startml

inconsistant axis label changes after using different trim metrics in plot

Plot is having issues after running trim with RMSLE