Is there an R package for Locally Interpretable Model Agnostic Explanations?

Question

One of the researchers, Marco Ribeiro, who developed this method of explaining how black box models make their decisions has developed a Python implementation of the algorithm available through Github, but has anyone developed a R package? If so, can you report on using it?

Answer 1

I think you're talking about the lime Python package. No, there is no R port for the package. The implementation for the localized model requires enhancements to the existing machine-learning code (explained in the paper), a new implementation for R would be very time consuming.

You may want to take a look at https://stackoverflow.com/questions/11716923/python-interface-for-r-programming-language for interfacing Python in R.

My suggestion is stick with Python. The package is only useful for highly complicated non-linear models, which Python offers better support than R.

Answer 2

Yes, there is now a port to R, which is available here: https://github.com/thomasp85/lime

It purports to provide LIME explanations for any classifier that implements a predict() method accepting a type = 'prob' argument. I have not yet tested it myself.