Talk:Sampling (computational modeling)

Undoing a deletion/ redirect of this page

User:MrOllie I just undid your redirect of this page, (you eliminated the page created by User_talk:Kozlova_Mariia with a redirect to the page Sampling (statistics)). My motivation is the following:

If one cares to read the two pages one will see that Sampling (statistics) is devoted to empirical experiments, either involving physical objects or individuals (humans) to be polled. This is about extracting entities from a population e.g. to set up an experiment in the laboratory or in a society as to ensure that several characteristics of the population are explored. Sampling in numerical experiments has to do with the exploration of multi-dimensional spaces to the effect of e.g. testing the output of a model, numerically integrating a function and so on. If anything, Sampling (computational modeling) is closer to Design of experiments (DOI) than it is to Sampling (statistics), though very few mathematical modellers use pure DOI but preferentially the methods in the newly created page. There is no conflict of interest in this page, neither mine not of the user creating it, and I consider that noticeably is ensured for this page. An incise, just to be clear: I know and appreciate the work of the user who created the page but the page does not contain self promotional material for either of us.

A litmus test of the argument for the difference between Sampling (statistics) and Sampling (computational modeling) is that neither Sampling (statistics) nor Design of experiments contains Low-discrepancy sequences, also known as quasirandom sequences or quasi random numbers that are a best practice in computer simulation. Note the existence of a more specific and technical page on Quasi-Monte Carlo method. The newly created page is a useful bridge for users interested in numerical experiments. I suggest that before deleting this page again the opinion of other editors is polled. Andrea Saltelli Saltean (talk) 11:35, 12 February 2024 (UTC).

The linked methods describe themselves as statistical methods and link to Sampling (statistics). This is clearly the same topic, just applied to a slightly different domain. We should not have two articles about the same topic, just as we don't have Samplling (medicine), Sampling (social science), etc. - MrOllie (talk) 12:28, 12 February 2024 (UTC)

• Sorry to contradict you User:MrOllie, but please note that Wikipedia has sampling (medicine)
• Wikipedia does not have sampling (social sciences) but has sampling (music) and sampling (signal processing) as well. Even sampling bias.
• As clearly noted above, in both sampling (medicine) and in the missing sampling (social sciences) that you take as an example one extracts samples from populations (of rats, drugs, chemicals, humans, treatments); bar discipline specific features, this can be covered in sampling (statistics). In sampling for numerical simulation or computational modeling one explores multidimensional space and this is not a slightly different domain. For example the concept of discrepancy - central to the field of sampling for numerical experiments - does not work for drugs and treatments. See the discussion of quasi-random sequences above.
• As I tried to explain, if redirecting (which I disagree with), this should be to design of experiments, not sampling (statistics).
• As I proposed, it would be useful to see what other editors think of this disagreement. Though I am not a great expert of Wikipedia procedures, a speedy deletion request from your side would have been preferable to a redirection, as this would have allowed a discussion with more editors.

Hoping that you will consider my reasons, I remove again your redirect. Best regards.Andrea Saltelli Saltean (talk) 11:44, 13 February 2024 (UTC)

Those aren't repeating the same material about statistical methods, as this article is. Speedy deletion requests are resolved without any discussion, that is why they are speedy. MrOllie (talk) 13:38, 13 February 2024 (UTC)