Skip to contents

The goal of NNTools is to facilitate the use of Machine Learning’ technics wrapping all the pipeline in easy to use functions (pipe_*). The functions take care of the cross-validation, data scaling and return the predictions (also for rasters), SHAP in a single call.

Pre-requisites

Install keras and tensorflow (only needed for pipe_keras and pipe_keras_timeseries):

keras::install_keras()

Installation

You can install the development version of NNTools like so:

# install.packages("remotes")
remotes::install_github("jmaspons/MLTools")

## The package "data.table" (>= 1.14.9) is required.
# data.table::update_dev_pkg() # requires unreleased features for reshaping 3D data