Wrap-up Quiz M6 Question 3

Module 6. Ensemble of models Wrap-up quiz
#1

In this question, which I adjoint below for completion, the result (even the result in the correction) is 0.5159872721018232 which is bigger than 0.5+0.04.
So the correct answer is Yes, however the marked is No.
Why? I am missing something?

Question 3 (1 point possible)
Replace the DummyClassifier by a sklearn.tree.DecisionTreeClassifier and check the statistical performance to answer the question below.

Is a single decision classifier better than a dummy classifier, by an increase of at least 0.04 of the balanced accuracy?
#2

@ miraunpajaro 0.5+0.04=0.54 > 0.516 :wink:
I did same mistake than you I read 0.004 and not 0.04 and I replied yes too :upside_down_face:

#3

0.04 or 0.004 doesn’t matter, DecissionTree is still better by those parameters

#4

the question is:

Is a single decision classifier better than a dummy classifier, by an increase of at least 0.04 of the balanced accuracy?

You obtained 0.5159872721018232 with the decision classifier and 0.5 as balanced accuracy with the dummy classifier.
0.516 - 0.5 = 0.016 and 0.016<0.04 so the answer is: no.
Do you get it better like that?

#5

As a matter of fact. I do

#6

:smiley: :+1:

#7

I think we need to look at all this wrap-up quiz in more details to improve it (other similar . I am tagging this so that we don’t forget to tackle it for the next MOOC session. Edit: actually this question is fine, I read the post too quickly …

@echidne thanks a lot for answering other user questions, this is really welcome and useful :pray:!