8
$\begingroup$

I just started wondering if there is anything that we can do with questions that are predestinated to become unnoticed and unanswered while having some potential.

Take as example this recent question: When have I to stop looking for a model? As I read it, it basically asks: is there anything that we can do in big data/machine learning scenario to find out that we are not going to learn anything more from our data. I'd say that this is an interesting question!

OP, prompted to edit, has tried to describe his problem in detail, but the question is pretty chaotic and I guess it may discourage potential answerers from reading it. I don't feel that upvoting it only to make it more visible would be a good way to go (it's not what upvotes are for). On another hand, making drastic changes so that it is more clear is not what edits are about.

$\endgroup$
3
  • 6
    $\begingroup$ If the question can't be readily edited into shape you can always ask a better question (and even answer it). It may be that the original closes as a duplicate (if not already closed for other reasons)... (but see my additional comment below) $\endgroup$
    – Glen_b
    Jul 4, 2016 at 17:22
  • 1
    $\begingroup$ @Glen_b I was thinking about the same solution but was unsure about creating a duplicate question... Posted it: stats.stackexchange.com/questions/222179/… $\endgroup$
    – Tim Mod
    Jul 5, 2016 at 8:23
  • 2
    $\begingroup$ if it would be close enough to close as a duplicate you should probably be editing instead. On the other hand, see this post which warns about overly heavy editing that risks changing the question. It can be a difficult course to steer between avoiding exact duplication and avoiding changing the sense of the question. $\endgroup$
    – Glen_b
    Jul 5, 2016 at 8:42

1 Answer 1

Reset to default
7
$\begingroup$

Posting my own, rephrased and more general, question helped: How to know that your machine learning problem is hopeless?. In the end my question received pretty big interest, number of up-votes and one very good answer. As a side effect, the initial question that I was referring to attracted much greater attention. So this case shows that posting my own question as @Glen_b suggested, was a good way to go.

$\endgroup$
1

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .