Issues with question 5

Hi;

I have issues with question 5. I got this question wrong and looked at the “show” answers. Running scikit-learn’s solution both in the sandbox version here (hopefully you guys have access to it) as well as on my computer, I do not get the same results for hgbdt than what’s shown in the “Show Answer”, as the correct result. There’s a huge differences in numbers, enough to skew the results of the answers to these questions.

Moreover; in this same notebook I’m showing a different answer to question 4, than what is showing as correct, using scikit-learn’s code for the solution.

Can you guys please look into this? I’ve saved my sandbox notebook, hopefully you guys can access it to check.

Cheers!

The answer to this question is quite stable through different folds. So my only guess is that the same issue with the definition of data I mentioned in your other post may be the origin of this problem.

Try running the exact code from all the solutions we provide up to question 5. In particular, try defining a different data matrix data_linear_model for the feature engineering part of the exercise, instead of adding columns to the original data.

Let us know if the problem persists after trying the above solution, as unfortunately we don’t have access to your notebooks.