The question is not explaining clearly what is needed.
I have done all the steps correct , and even the explanati0on of answer is not matching with the answer c
1 Like
The question asks you to compare two models fold by fold, which is then demoed in the answer. What would you say is the troublesome part of the question?
In my case I get:
The model using all features is performing better 3 times out of 10 than the model using only numerical features.
But probably I have done something wrong. It does not matter, I am doing the mooc just for fun, and learning not for a mark which probably has no meaning in the “real” world.
Thanks
I did not understand for example what do you mean by [7,10]
The square brackets denote a closed interval i.e. an interval which includes all its limit points. In this case, [7,10] means greater than or equal to 7 and less than or equal to 10. Using this same example, an equivalent form to read option c) would be:
[7,10]: the model using all the features is better than the model using numerical features only in at least 7 out of the 10 folds.
Even if you are not really interested in the note, if you want to share a piece of your code maybe we can help you debug it. The model with all features should definitely perform better as it has more information fed in.
We should probably rephrase this “between 7 and 10 folds (included)” in a future version of this mooc.
1 Like
Hi,
I found very different results (best performance only 5 times) so I redid with the code given in the explanation and I find exactly the same results … (clear kernel and re- load the dataset and lauch the explanation code).
I tagged this as mooc-v4. I think we could even list all the values i.e. 7, 8, 9 or 10 folds
Yes, that’d be much clearer.