2
$\begingroup$

I am really puzzled that this question received no response, it does not seem to be a difficult question, it is simply asking for the opinion of when to use one method over another. Is this because it is a dumb question (very possibly), or hard to answer (perhaps in depth, but I would be grateful for any insights), or is it somehow off topic?

ARIMA vs. Random Forest

I just want to know because I really thought this was the place to ask it.

$\endgroup$
1
  • 2
    $\begingroup$ I can only speak only for myself -- I simply think other people than myself would be better placed to respond that. You may want to consider whether more details, some clarification, or greater visibility might help. $\endgroup$
    – Glen_b
    May 3, 2015 at 13:47

2 Answers 2

Reset to default
5
$\begingroup$

It's hard to read. I know it's not a long paragraph, but something about reading in a web browser makes paragraphs of even modest length feel like an eternity. I always use lots and lots of line breaks when I post here and I appreciate it when others do the same.

Or just look at any post by one of the top users. You very rarely see paragraphs bigger than you'd find in a newspaper.

It seems trivial but it makes a huge difference. There is a relatively small (and possibly shrinking) number of active, knowledgeable users on this site. Even a tiny difficulty in answering a question can be a good excuse to skip it. That's especially true when, as gung said, the topic is specialized and chances are any answer will involve some research.

Also, "power load functions" are not something most people will be familiar with. You absolutely must describe your data.

And finally, I hate to say it but it's also kind of a "let me Google that for you" question. I know because I Googled "time series random forest" and I immediately got 5 results that seemed like they'd be directly helpful. I'll post them in a comment on the question itself.

$\endgroup$
1
  • 1
    $\begingroup$ (+1) Don't think it's hard to read, but the others are shrewd observations. (1) Inclusion of untranslated jargon can make the problem appear potentially more complicated than it is - do we have a different function of something at each time point to consider, or just a value of a function? As well as being tedious to have to ask, it suggests an unthoughtful questioner, which doesn't bode well. (2) Doing some research beforehand can allow you to focus the q. & make it more enticing to people who might answer. $\endgroup$
    – Scortchi - Reinstate Monica Mod
    May 4, 2015 at 14:10
2
$\begingroup$

Statistics has more nooks and crannies than a Thomas' English Muffin. Knowing how well you can forecast with an ARIMA vs. a random forest strikes me as very specialized. It may be that there are few people who feel sufficiently expert in both areas to comment.

I also notice that your question has had rather few views. You might try something to get the question to attract more attention overall. It may be the few people who could answer haven't even seen it.

$\endgroup$

You must log in to answer this question.

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