Wikipedia:WikiProject Mathematics/A-class rating/Maximum spacing estimation/Archive 1

The following discussion is preserved as an archive. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Maximum spacing estimation edit

Closed due to inactivity and some lingering concerns. Last time the article was edited is March 6. There is no prejudice against another nomination in the near future if someone wants to work on it. My summary of the discussion below is that I think in general there are some accessibility concerns. While I don't think the A-class criteria forbids topics not very accessible to laymen, some reasonable attempts should be made at readability. This may include reiteration of some very basic stuff in brief form and with simple examples. --C S (talk) 02:54, 30 April 2009 (UTC)[reply]

“And what is the use of a book,” thought Alice ”without pictures or conversation?” I'm not sure exactly what one could do to illustrate the concepts in this article, but a picture or two might help make it a little less dry. An example would help, too. The article violates WP:LEDE in that there's information in the lede that's not elaborated elsewhere (the history of the method's discovery): “Significant information should not appear in the lead if it is not covered in the remainder of the article”. Additionally this article lacks context (what other methods solve the same problems and under what circumstances would one prefer one method to another). So it seems a good start but somewhat shy of the A class to me. —David Eppstein (talk) 02:28, 5 February 2009 (UTC)[reply]
1. I'm not exactly sure what kind of picture to add, unless you want a q-q plot of a fit vs. empirical data or something. -- Avi (talk) 03:12, 6 February 2009 (UTC)[reply]
2. Ditto for examples. -- Avi (talk) 03:12, 6 February 2009 (UTC)[reply]
3. I can move the history into the first section after the lead, if that is better. -- Avi (talk) 03:12, 6 February 2009 (UTC)[reply]
4. The other method commonly used, maximum likelihood is mentioned in the article, and the reason for using MSE is stated more than once, IIRC. What expansion would you feel is appropriate? -- Avi (talk) 03:12, 6 February 2009 (UTC)[reply]

- - Seems strange to say you don't know what examples to add. I'm sure someone teaching a class or writing a textbook would be able to think of some. I think the reason for using MSE is not at all clear. What are the pros and cons of this method versus other methods that are used when maximum likelihood is not applicable? Spell it out. What are specific situations when unambiguously MSE would be the better method to use? --C S (talk) 22:49, 7 February 2009 (UTC)[reply]
    - An example would be a stream of input and the output, pretty boring, but a simple example can be added. -- Avi (talk) 09:04, 8 February 2009 (UTC)[reply]

Oppose per David. (not part of the project, but trying to help review) —Ed 17 ^{(Talk / Contribs)} 21:21, 5 February 2009 (UTC)[reply]
- Not being part of the project, you probably are not aware of the norms here. This is not a vote, but a discussion. "The "support"/"oppose" approach to review should be avoided as far as possible." from Wikipedia:WikiProject_Mathematics/A-class_rating; see also Goal 3. --C S (talk) 22:45, 7 February 2009 (UTC)[reply]

Comment. It doesn't seem like the article is following the math manual of style. This is evident from the very first sentence. In addition, there is not sufficient motivation for the topic. I'm still unclear as to why anybody studies this other than they can. Why is the fact that it is consistent with Kullback–Leibler information theory important? Why is it discussed in "various applied science" journals? While it is not explicit in the A-class guidelines, it is implicit in the language that a nonmathematician be able to get something out of it. This is explicit in the math MoS, which A-class articles should attempt to follow. The MoS suggests examples and in informal introduction. I think this would go a long way in this article. The article seems to consist of mostly a formal definition and list of properties. --C S (talk) 06:16, 7 February 2009 (UTC)[reply]
- Sorry about the delay; work is heating up. Here are responses to your excellent points:
  - Consistency with Kullback–Leibler information theory is important because it places MSE in the same fundamental framework as the better known MLE, and also demonstrates that MSE is a method which can be conceived as converging in some fashion to the "true" model ala Anderson and Burnham (who discuss MLE, but the theory is the same).
  - The "various journals" sentence was to show that this method has been discussed and used outside of purely mathematical and statistical fields.
  - The method is studied as an alternate to the current workhorse of statistical inference, MLE. As MLE has certain specific weaknesses, and MSE is more more robust in these specific areas, MSE is worth study, and more importantly, actual use in statistical inference.
- -- Avi (talk) 04:23, 16 February 2009 (UTC)[reply]
Any replies or new comments? -- Avi (talk) 15:25, 1 March 2009 (UTC)[reply]
- Just some quick comments/questions for now (forgive the barrage of questions): it doesn't seem as if the article has been improved according to the above suggestions since the review started. It's fine to put clarifying remarks here, but I would think it would be better to put them into the article itself. Also, someone commented on the talk page that MSE isn't usually used by itself. I think this is worth clarifying. The article still is unclear on why this method is useful. What are the examples of when it is better? What are these specific weaknesses and areas you mention? Why aren't they in the article? Can any simple examples be added to illustrate some of the concepts? Why is there no mention of how MSE compares to alternate methods? A lot of problems would be alleviated if some attempt was made to follow the math MoS, which suggests examples and an informal suggestion. --C S (talk) 00:20, 2 March 2009 (UTC)[reply]

I'm going to do a little copyediting, but I'm not familiar so I can't do very much. I do think that a more elementary presentation would be preferable, at least in the lede. — Carl (CBM · talk)
P.S. After going through the article again, I have a few additional suggestions. (1) We ought to have a short section that sketches the applications that are alluded to in the lede. (2) Perhaps we could include a very simple example, to illustrate the method more concretely. — Carl (CBM · talk) 00:52, 6 March 2009 (UTC)[reply]
- Good points. I'm sure I can whip something simple up, although it would be boring :) -- Avi (talk) 00:57, 6 March 2009 (UTC)[reply]

The above discussion is preserved as an archive. Please do not modify it. Subsequent comments should be made on the appropriate discussion page, such as the current discussion page. No further edits should be made to this discussion.