๐Ÿš€ R Starter Kit

Download Execute In Colab
Notebook
1 Like

Would it be possible to provide some further details on how to proceed to a proper submission?
Despite having passed the tests and and getting a messing acknowledging a proper submission, the leaderboard evaluation returns errors with little indications on where the error is coming from.

For example, I first wanted to use data.table library, despite having the following to work properly in the Colab notebook, it resulted in the error message "Error in data.table(x_raw) : could not find function \"data.table\"" on the submission evaluation section.

install_packages <- function() {
  install.packages("data.table")
    require("data.table")
}

Looking at the submissions, it seems like others also faced issues package installation/usage, so I think some further examples would be welcome.

Also, when opting for the zip submission instead of the Colab, it resulted in "DockerPushError: An image does not exist locally with the tag: aicrowd/imperial-pricing-game". The submission experience has been a little rough so far on the R side!

2 Likes

Iโ€™m getting the following error when running the default notebook for R:

1] โ€œIncome: 26057988.08โ€
[1] โ€œLosses: 26057988.08โ€
[1] โ€œYour model is invalid: it loses money on its training data!โ€

Hello @simon_coulombe

Itโ€™s not an error. You can submit the model and you will get a score. The message means that the performance of the model is not good.

1 Like

thanks for replying!
Iโ€™m just surprised that it says the model is losing money on the training data since the income is equal to the losses. The rules do state that you are not allowed to submit a model that loses money on the training data so I was afraid that might be an issue down the road.

Hi @simon_coulombe

Yes indeed this is a good point. Two things two consider for now:

  1. We are currently implementing a warning system upon submission on the AICrowd website so that you are warned if your submission is not profitable on training data. If it is not the model will not be entered into the profit leaderboard so itโ€™s important to know in advance. Stay tuned for this update within the next day!

  2. The rule requires positive profits. So a profit of zero is technically invalid. Hence the warning you received :slight_smile:

1 Like