Talk:Change detection

Direction

I think we must mention non-statistical uses if we do not want the article to merely duplicate sequential analysis.--Adoniscik (talk) 01:12, 13 January 2008 (UTC)

Revisions

From my perspective, the article needs some important revisions. Up to now, there are several contents without any sources (e.g., "In a Bayes change-detection problem, a prior distribution is available for the change time", "Online change detection is also done using streaming algorithms", "A key technique for minimax change detection is the CUSUM procedure.", "Change detection tests are often used in manufacturing, [...], website tracking, and medical diagnostics"). Moreover, several sections are not concise and complete (Minimax change detection, offline change detection, applications of change detection). Content (with sources) is missing such that the article is not informative and useful up to now. Thus, it is clear that the article is rated as "Start-Class" and "Low-Importance". Content and sources are missing. Finally, "Linguistic change detection" belongs to the section "Applications of change detection" from my point of view.

My proposal: A) Reordering of the sections including subsections: 1 Types of change detection, 1.1 Online change detection, 1.2 Offline change detection, 1.3 Minimax change detection, 2. Applications of change detections, 2.1 Applications of change detection tests, 2.2 Linguistic change detection. B) Adding important content: The EWMA procedure is missing in the section "Minimax change detection". In section "offline change detection", other clustering algorithms should be mentioned. Moreover, applications should be explained in more detail (incl. providing the respective citable sources), like in section "Linguistic change detection". That generates content to make this section useful. — Preceding unsigned comment added by PhilOt (talk • contribs) 12:13, 22 October 2016 (UTC)

As for reordering, I don't know. The article does look like it can be improved. But we do not need a citation on every sentence and example but only for "likely to be challenged" claims. In particular, not a 2016 source that is far from well-known but only cited once, by the author himself. If you want to add references to statements, please check first if the Basseville textbook is suitable. That book may also provide a guideline of how to structure the article. Because as far as I can tell, this book a respected source, while your reference is not. Also I do not think applications need to be explained in detail; this is an encyclopedia, not a survey. Example sections are usually just advertisement for some crappy papers; and nobody reads that part. One or two words each, with no references, is enough to know it can be done. You are welcome to improve the text of the article, but don't add dubious sources again, thank you. Also, don't just try to come up with an excuse to re-add the reference. I doubt it is "likely to be challenged" that change detection has been used in medicine; so no need to reference that (and again, preferably find a highly cited book that lists this use case)! Adding references like your does not increase the article quality, you know. HelpUsStopSpam (talk) 19:47, 22 October 2016 (UTC)

It is clear that it is not needed to have citations on each sentence. However, I really doubt that it is obvious that change detection procedures are used in medical DIAGNOSTICS. The application of change detection procedures in spatial statistics is "likely to be challenged", so the reference I have cited was definitely needed. Moreover, I absolutely disagree that there is no need of sections on applications and that nobody reads these parts. If you have a look on the high-quality articles (WP:STATS) on statistical models, you will see that various applications are explained for almost all of them. Moreover, the added reference you called "dubious source" is a reliable, published source (WP:SCHOLARSHIP). In this context, note also WP:MEDDATE PhilOt (talk • contribs) 17:03, 15 November 2016 (UTC) — Preceding unsigned comment added by 193.174.122.17 (talk)

Clearly, your only interest is dropping this one reference on Wikipedia rather than actually improving the article. As a matter of fact, Google Scholar has many earlier mentions of change point detection in medical diagnosis. In fact, the 1993 Basseville book I mentioned above has a section with examples, including "11.2.2 Biomedical Signal Processing" giving several sources from the 1970s and 1980s, or e.g. this from 1999: http://www.sciencedirect.com/science/article/pii/S1361841599800012 . Your have a conflict of interest here: you want your publication to be cited, not the most relevant publications... stop advertising. HelpUsStopSpam (talk) 21:52, 15 November 2016 (UTC)

There is no conflict of interest according to WP:COI. PhilOt (talk • contribs) 09:18, 16 November 2016 (UTC)

Then why is the IP you edited with above at Viadrina European University, where the authors of that one reference are? Sorry, I have good reasons to assume that you have a COI and at least know one of the authors personally and thus may be biased and overestimate the relevancy. Even if you claim to not have a COI there are good reasons to believe otherwise. And your behavior also clearly shows that you don't appear to be really interested in improving the article, but are obsessed with getting that one marginally relevant reference into Wikipedia... HelpUsStopSpam (talk) 21:15, 16 November 2016 (UTC)

Wiki Education assignment: Human Cognition SP23

This article was the subject of a Wiki Education Foundation-supported course assignment, between 20 January 2023 and 15 May 2023. Further details are available on the course page. Student editor(s): AbigailG23, Dhermatology (article contribs).

— Assignment last updated by Dhermatology (talk) 19:53, 13 April 2023 (UTC)