Timeline for Approaches for generating synthetic survey data with dependent answers?
Current License: CC BY-SA 3.0
29 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 13, 2017 at 12:44 | history | edited | CommunityBot |
replaced http://stats.stackexchange.com/ with https://stats.stackexchange.com/
|
|
| Mar 18, 2015 at 20:43 | vote | accept | Vass | ||
| Mar 16, 2015 at 23:32 | answer | added | Glen_b | timeline score: 5 | |
| Mar 16, 2015 at 22:55 | history | reopened |
Glen_b whuber♦ |
||
| Mar 16, 2015 at 15:12 | history | edited | Glen_b | CC BY-SA 3.0 |
added 47 characters in body
|
| Mar 16, 2015 at 2:30 | comment | added | Glen_b | Okay, I reconstructed something very similar (and saved all the text and code this time). If your post reopens I'll paste an answer. | |
| Mar 16, 2015 at 2:07 | comment | added | Glen_b | Vass Much improved. Unfortunately, my computer crashed in the meantime, so the detailed example that I had sitting waiting to paste in as an answer a couple of days ago is gone since I last commented. I can try to do another I guess, but I'm less inclined to go to the same amount of effort. If I'd anticipated the time-gap (which perhaps I should have), I'd have saved a copy of it. | |
| Mar 16, 2015 at 1:37 | review | Reopen votes | |||
| Mar 16, 2015 at 9:11 | |||||
| Mar 16, 2015 at 1:22 | comment | added | Vass | @Glen_b, I made edits to the question which I hope is more suitable. | |
| Mar 16, 2015 at 1:21 | history | edited | Vass | CC BY-SA 3.0 |
added 376 characters in body
|
| Mar 14, 2015 at 23:19 | comment | added | Glen_b | The example I did has 3, 4 and 5 categories for the three variables but you can see how to change it I think. It should also be clear how to generalize to some veriables being non-categorical, and yes, MI should see it. Can you edit your question to ask one or two more specific, detailed, questions (i.e. with details like you just gave -- which would invite Antoine's answer, and which would invite a categorical answer). A more specific question should re-open and I can then post. | |
| Mar 14, 2015 at 16:47 | comment | added | Vass | @Glen_b, I don't use R, but can understand it. From a short example I can then edit my answer to be worded in a way that is not too broad. I need any type of dependency which would show up when measuring the mutual information between question answers. Category numbers 2 or 3? | |
| Mar 14, 2015 at 1:39 | comment | added | Glen_b | Vass - how many categories do you need? What sort of dependence would you like? I could show some examples if your post comes off hold (I can't post an answer while it's on hold), but they might not be what you want. I have a very short example coded up in R that might be helpful if you use R. | |
| Mar 13, 2015 at 16:49 | comment | added | Vass | @Glen_b, let's say I generate randomly and independently answers to 2 different questions with categorical questions. I want to generate an answer to a third question which depends on the first two answers. The answer by Antoine pretty much explains how to do the cintinuous valued case, but I am still unsure how to do the discrete case, could you tell me how to do that? | |
| Mar 11, 2015 at 21:36 | vote | accept | Vass | ||
| Mar 18, 2015 at 20:43 | |||||
| Mar 11, 2015 at 17:23 | history | closed |
Glen_b Xi'an gung - Reinstate Monica Scortchi♦ Tim |
Needs more focus | |
| Mar 11, 2015 at 15:12 | comment | added | Glen_b | This question has suggestons for basic reading on copulas | |
| Mar 11, 2015 at 15:07 | review | Close votes | |||
| Mar 11, 2015 at 17:23 | |||||
| Mar 11, 2015 at 15:03 | comment | added | Glen_b | Your question is not clear enough about what you really need for me to formulate your questions for you, but I really don't think you need the whole general theory; something like mixtures combined with item response theory might well be sufficient for 99% of what you need. Indeed something even simpler might well do. | |
| Mar 11, 2015 at 15:01 | comment | added | Glen_b | One simple (but not remotely general) way to introduce dependence in ordinal categories is with latent variables. You might look at item response theory, for example, as giving a class of models from which dependent categorical data could be generated. | |
| Mar 11, 2015 at 14:49 | comment | added | Vass | @Glen_b, I tried reading the wiki article on copulas, it will need some investment to understand deeply. The Nelson book on it is 55 for the kindle, ouch. Maybe I can ask 3 new questions related to the subject if they already do not exist? How would you phrase them to get good answers? | |
| Mar 11, 2015 at 14:45 | comment | added | Glen_b | (ctd) ... finally, you're going to have to be much more explicit about what you mean by "test", but my guess is that would also tend to be long for an answer. Ask one specific question please. But I suggest you tackle the wikipedia article first, and then some of the readings it suggests, and then some of the copula questions here, in order to be able to ask a suitable question. The alternative is to drop the insistence on generality and tackle a smaller problem. This is not something that can be conveyed easily in minutes. | |
| Mar 11, 2015 at 14:45 | comment | added | Glen_b | You seem to misunderstand the scale of what you're asking for. Enough of the general theory can't be usefully conveyed in a paragraph without being facile to the point of uselessness (see the opening 3 paragraphs of the article I linked to). If your primary interest is categorical variables there are additional complications arising from the general theory, so you wouldn't want that in your first example, and even if you stated what sort of dependence structure you wanted, even one example is likely to be getting on the long side if you need it at a basic level, and ... (ctd) | |
| Mar 11, 2015 at 14:34 | comment | added | Vass | @Glen_b, since I do not know enough to be able to guide the answer to fit a page, how about a paragraph on the general theory to understand where the problem lies and the ability to search for more like with your great link to copulas, an example or two for nominal and ordinal variables, and some ways on how to test the results? | |
| Mar 11, 2015 at 14:25 | comment | added | Glen_b | The general theory I've already mentioned. It's a subject that fills books; I have several of them on my shelf - could you perhaps ask a question that might be answered in half a page or so? | |
| Mar 11, 2015 at 14:19 | comment | added | Vass | @Glen_b, I would like to start off simple, and then make it more complicated over iterations. I would like to know the general theory behind introducing dependencies, and different ways of creating dependencies. And then how to measure the dependencies created. | |
| Mar 11, 2015 at 14:14 | comment | added | Glen_b | The general approach to inducing dependence would be via copulas, but that may well be overkill for your problem. If you give more details about what you need, something simpler may well be feasible. | |
| Mar 11, 2015 at 13:02 | answer | added | Antoine R | timeline score: 4 | |
| Mar 11, 2015 at 11:27 | history | asked | Vass | CC BY-SA 3.0 |