1

Recently I started to do some researchs on QA data aiming to rank quality answers from determined questions, for this I used the avaliables Yahoo! and StackExchange datasets.

Now, I would like to expand the research for identify potential answers to the thread starter along all posts on that thread. For this, I searched for some online forum dataset, possibly annotated (like the Ancestry.com, unavailable now), because on online forum a reply on a thread not always means that is a direct answer to the thread starter. But I couldn't find any dataset like this.

Do you guys know about some avaliable dataset that could help me? Thanks.

2

The StackExchange sites themselves seem your best bet!
(Note: Question was edited after this to exclude StackExchange as well)

Read Introducing Stack Exchange Data Explorer, which is about SEDE, where you can run queries, and about the dumps of all the public data in the current Stack Exchange network.

This means you can either access the data through SEDE or download it all and play with it, with proper attribution (Creative Commons Attribution – Share Alike) of course.

  • Thanks for the reply! I also realize that the StackExchange sites could help and I have done some tests with good results, but unfortunately this dataset has the same problem of Yahoo!, I can only rank the quality and can't identify what could or not be a candidate answer, because all of them are, almost always, direct candidate answers to the inicial question, different of the posts along a forum thread =( – Max Willian Oct 4 '17 at 21:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.