Now that we have some collaborative software development happening thanks to @matteo-fontana @paolo-vergo and team, we should really implement unit tests in this package. We can use the testthat package, for example.

Paolo and team, can you guys start doing this at least for the new stuff you're adding? (Ideally we'd write tests to cover everything that's there as well---if you wanted to do that, then this would be excellent.)

Hi @ryantibs
I'm trying to use the model on a different data set.
It's multivariate regression problem.
A matrix X to predict a y value.
A highly nonlinear function.
What would be the best strategy using the package?

Many thanks in advance for your assistance.
Best,
Andrew

The package in its current form is not uploaded on CRAN. It would be useful to have it there, so to increase the visibility of the package

It would be useful to implement, as a prediciton method, the Jaccknife+ by Barber and coauthors (https://arxiv.org/abs/1905.02928)

Now that @matteo-fontana @paolo-vergo and team are planning to add a bunch of new functionality, we should:

• update the authors in the DESCRIPTION to include the new folks who contribute code
• update the README to include the new folks as well, as well as the new papers whose methods are reflected in the code
• move all the documentation over to pkgdown/github docs. This (in my opinion) is way more readable and nice than the "old style" pdf documentation file.
• assuming we do the last one, we should also move all the examples over to vignettes in the pkgdown site. Again, this (in my opinion) is a way nicer way to see examples of how to use the package.

Hi,
since conformal prediction also works for classification, do you plan to implement it for classification here in the near future?

Best regards,

Robin

The current implementation of split conformal regression supports only equal splitting: it would be useful to add the possibility of explict specification of the share of observation to be in the training set.

It would be useful to implement, as a prediction method, the Multisplit Conformal one by Solari & Djordjilović (https://arxiv.org/pdf/2103.00627.pdf)

When running examples from the r code folder, examples such as lasso.funs and elastics.funs, I keep getting an error when running it.

funs = lasso.funs(lambda = lambda)
Error in !is.numeric(lambda) || min(lambda) <= 0 || (order(lambda) != :
'length = 96' in coercion to 'logical(1)'

This is the error I am getting when running these functions. Please help out if possible.

It would be interesting to implement Conformalised Quantile Regression (Romano et al 2019) (https://arxiv.org/abs/1905.03222).
This could be done via the implementation of other Non-Conformity Measures