Help identifying forum posts

[art297]

I'm currently working on a project where everyone's expertise would be of enormous help.

To date, together, we've contributed more than 80,000 posts to the forum.  I am currently mining those posts to come up with a framework that, hopefully, will make it easier for everyone to tap those posts in order to quickly find the answers to their questions if those questions have already been answered on the forum.

That is where you come in.

Please send me as many examples as you can.  For each example, I need a statement of the question that was addressed, and links to posts that provided good answers to the question.

You can either post them as replies to this question, or send them to me, directly, at art297@rogers.com.

Thanks in advance,

Art

[Linlin]

Hi Art,

I just emailed you the updated file. The forum doesn't allow me to attach here.

BTW, is the second turkey successful?  - Linlin

[art297]

Other than a response from LinLin, I haven't received ANY other responses.  However, I just discovered that art297@rogers.com was blocking most email addresses.  I think the problem has been resoved but, just to be sure, please send your links to atabachneck@gmail.com

TIA,

Art

[TomKari]

: I just resent two to your gmail address, so you can see if they come through.

Tom

[art297]

: They did and very much appreciated!

[Doc_Duke]

Art,

Your question is nearly impossible for many of us to answer.  When I have a question, I just google it

<topic...> site:sas.com

and get the things that SAS has written intermixed with the ones in the forum.  When I get the answer, I go on; I don't record it so I don't have anything to send to you.

Doc Muhlbaier

Duke

[art297]

: For those of us who know what we are looking for, and know where to limit the search, that method can often work.  However, sometimes the answer isn't found at support.sas.com, but may be on one of the other SAS related venues or even just lost somewhere out on the web in a user's site.  And sometimes, even for those of us who are quite experienced at finding such answers, the answer might end up being embedded with 1 or 2 million hits that aren't anything close to what we were looking for.

As I mentioned to KSharp in a carbon copy of this thread: TextMiner's algorithm can statistically identify clusters of documents by itself, but the clustering can be enhanced with models that can be obtained from Content Categorization Studio.  Those models, in turn, are developed by stipulating a category (i.e., a question like "how can I obtain the previous and next record") and running the program using up to 20 documents/posts/blog entries that experts have identified as representing a correct answer to the question.

For this purpose, all SAS users are experts

What I'm looking for from SAS users, is simply a question and a link to a post/web page/blog entry/paper that provides a correct answer to the question.  Obviously, the less noise included in the link the better (e.g., the full SAS reference manual would NOT be as good of a source as a post that specifically provides a direct answer to the question, and a thread that contains a post that answers a question would not be as good as the one post that really does provide the answer).

So, what I'm looking from experts like you, is a list of questions (one or more) and, for each question, a link that provides an answer to the question.  Even if one doesn't maintain such a list, sending some such examples as they are found, would work too.