Wikipedia:Articles for deletion/Mayfly optimization algorithm

The following discussion is an archived debate of the proposed deletion of the article below. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.

The result was redirect to List of metaphor-based metaheuristics#A mayfly optimization algorithm (MA) (Zervoudakis & Tsafarakis, 2020). No experienced editor wants to keep this, and the proposed redirect is an appropriate WP:ATD. Sandstein 16:16, 20 April 2022 (UTC)[reply]

Mayfly optimization algorithm

[Hide this box] New to Articles for deletion (AfD)? Read these primers!

(Find sources: Google (books · news · scholar · free images · WP refs) · FENS · JSTOR · TWL)

However noteworthy this may be, it doesn't seem to have been much noted. NN.

Yet links to it -- and to "Flying Fox optimization" aka "Flying Fox Optimization Algorithm", another creation of Konstantinos Zervoudakis and Stelios Tsafarakis -- have been added rather vigorously, to "Genetic algorithm", to "List of metaphor-based metaheuristics", to "Table of metaheuristics", to "Heuristic (computer science)", and to "Differential evolution".

The article was created by someone with a declared COI, and the recent vigorous linking to it seems compatible with undeclared CoI. -- Hoary (talk) 03:31, 12 April 2022 (UTC)[reply]

Note: This discussion has been included in the list of Mathematics-related deletion discussions. Hoary (talk) 03:34, 12 April 2022 (UTC)[reply]

Delete. Bubba73 ^{You talkin' to me?} 03:57, 12 April 2022 (UTC)[reply]
Comment. The main paper on this topic has 127 citations already, as listed in Google Scholar, high for a 2020 publication. Nevertheless I strongly believe that this whole line of biologically inspired optimization algorithms, despite some initial successes, has devolved to become largely or almost entirely junk, or maybe pyramid-scheme science (if you hype up your topic enough to bring in enough new researchers to cite your papers, you can get high citation counts and succeed in measures of quality and performance based on citation count regardless of the boilerplate nature of the actual research). For a more informed insider opinion reaching similar conclusions, see "Metaphor-based metaheuristics, a call for action: the elephant in the room", doi:10.1007/s11721-021-00202-9. The question for us here, though, is, can Wikipedia guidelines justify removing content because we believe it to be junk, or must we keep popular junk? —David Eppstein (talk) 06:59, 12 April 2022 (UTC)[reply]

Metacomment: "Metaphor-based metaheuristics, a call for action: the elephant in the room" is open access, and very readable. -- Hoary (talk) 09:15, 12 April 2022 (UTC)[reply]

A suggestion to delete an article from wikipedia that lacks logical or scientific evidence with only the valid argument the reference to the idea of the article "Metaphor-based metaheuristics, a call for action: the elephant in the room" lacks seriousness and I cannot characterize it something other than funny. As I imagine the author of the comment has not understood what the article he quoted mentions, I beg him to consider in which journal the mayfly algorithm is published, what is the impact factor of the journal, who are the reviewers of the journal and many other similar information that we all have in our mind when judging a scientific paper. So how is it concluded that an article say that algorithms of this type are junk? Does this article focus on the features of published unreliable algorithms? Do not try to create what is called in philosophy a “logical fallacy”. If you want, prove your claims.

My view is to put aside our subjective judgments and our ideology of what is junk and what is not and let those who work on the field to judge it. With proofs. Besides, I do not think it results from the paper's references, several of which are in internationally renowned journals and from scientists in the field stating its validity, that as an algorithm it is usefulness. That is why you should let time judge it and not try to create impressions. Useless to me are the comments that try to persuade without any serious counter-argument, depriving the freedom of speech of its creators. Kostas 0231 (talk) 16:06, 12 April 2022 (UTC)[reply]

Those 127 citations include MDPI journals and physics conference proceedings, neither of which are known for a high standard of review. XOR'easter (talk) 16:31, 12 April 2022 (UTC)[reply]

Not only... The most are Elsevier, Wiley, Springer, IEEE... I see also journals with high impact factor... Do you think you have the authority to jurge and downgrade that? Beyond that, my purpose is not to prove something you disagree with. My goal is to remind you of the role of wikipedia. However, it is not censorship or ill-intentioned and unsubstantiated criticism. Kostas 0231 (talk) 16:59, 12 April 2022 (UTC)[reply]

I am, in fact, entirely comfortable with dismissing Impact Factor as a means of evaluating anything. My present opinions about the topic at hand broadly align with those expressed by Elemimele below, who aptly summarizes the "role of Wikipedia". XOR'easter (talk) 19:29, 12 April 2022 (UTC)[reply]

Something from Wikipedia and i stop here because several dont understand what freedom of speech means... "Wikipedia is an online free-content encyclopedia helping to create a world where everyone can freely share and access all available knowledge".

Maybe are better articles on junk diets, junk dietitians/nutritionists, junk myths, junk music, junk movies, junk sects, junk politicians, et cetera. Have a nice day. Kostas 0231 (talk) 20:52, 12 April 2022 (UTC)[reply]

Per policy, Wikipedia is free and open, but restricts both freedom and openness where they interfere with creating an encyclopedia. Accordingly, Wikipedia is not a forum for unregulated free speech. XOR'easter (talk) 22:20, 12 April 2022 (UTC)[reply]

Comment, David Eppstein we have to cover junk, if it's notable junk. We have to assume that if it's that bad, eventually secondary reviewers will start to say so, and then we say so. But you're quite right that some areas of science have such high hype-value that every paper is instantly cited by everyone else creating similar stuff, and the citation count can be very high, giving a misleading indication of notability. We have to remember that things like citation count are just proxies for notability that we've chosen to use, by convention. The bottom line is that the algorithm becomes notable when it gets solid independent secondary discussion, in reasonable depth. Many of these citations will be at best passing references, or bits thrown into primary literature merely to acknowledge the existence of an algorithm related to a different piece of work that the source's authors are trying to present, so as to convince a journal reviewer of their knowledge of the field. Elemimele (talk) 07:46, 12 April 2022 (UTC)[reply]
Yes, for example, Karami et al. (2021) just gives it a passing mention in a long sentence listing various algorithms, including the "wingsuit flying search algorithm". XOR'easter (talk) 16:41, 12 April 2022 (UTC)[reply]
Comment We have articles on junk diets, junk dietitians/nutritionists, junk myths, junk music, junk movies, junk sects, junk politicians, et cetera. Holding my nose, I say that other things being equal, junkiness is fine. Popularity isn't a criterion for inclusion, although it tends to determine written-about-ness, which determines "notability", which is a criterion. If a metaphor-based metaheuristic (or whatever) becomes "notable" (as defined by Wikipedia) then its exposure (whether simultaneous or subsequent) as worthless doesn't diminish its "notability". However, "notability" needn't, and here shouldn't, be simply determined by an easily grasped (but notoriously inflated) number that Google puts atop a page of search results. -- Hoary (talk) 09:15, 12 April 2022 (UTC)[reply]

I think that you have totally misunderstood the complete nature of an academic paper. And I can't accept the term "junk" about an algorithm that has been published on a high impact factor journal and used as the main algorithm in various research and was reviewed by well known researchers in the field.

This is really offensive.

It also appears that you searched about the 127 citations but you totally ignored that too many of those papers use this algorithm as the main algorithm of the research.

This is also offensive.

Moreover it feels like the main problem is the popularity of a method, which feels wrong.

I will add the papers that use the mayfly algorithm in the wikipage the following days.

As regards the paper that you mentioned regarding the quality of an algorithm, the mayfly algorithm was built, compared and tested according to this paper and according to various other metaheuristic-introduction papers.

I need about one week to improve the page. GusRDRM (talk) 14:07, 12 April 2022 (UTC)[reply]

Please don't feel personally attacked! It's not really like that. The situation is this: We mustn't forget that Wikipedia is a general encyclopaedia read by a very broad readership, it's not a primary, or even secondary source of academic information. Something can actually be very good, but it's not suitable for a Wikipedia article until it's been widely written-about by secondary sources. That may feel like "popularity", but it's really about ensuring we have reliable information before we start writing articles. The ideal point at which a WP article can be written about an algorithm is when the algorithm is mature enough to feature in textbooks and broad reviews about its applications, strengths and weaknesses etc.; not when it's reached the stage that it's being used in primary publications by people who have read the original primary publication. My impression at the moment is that the citations are largely people recognising that it exists, or applying it, not reviewing it. If you are improving the article, GusRDRM, I would advise you to find reviews by unconnected authors, if at all possible. And if you can find them, they will greatly contribute to the strength of the article, and its likelihood of surviving this AfD discussion. My comments about junk, above, were similar to those of Hoary, and really meant that it is not necessary to assess whether this algorithm is junk or good in order to assess whether it should have an article. It gets an article if someone writes about it. If it is later found to be junk it still gets an article, but we report, fairly, what's been said about it and its field. We have no axe to grind, we are not here to judge quality directly, only to judge our sources and reflect them accurately.

Summary: Raw citation count is a rubbish metric; you are quite right that actual usage matters more than a mere count. To my mind, there are three levels here: primary papers that mention the algorithm but don't discuss it in depth or use it (citations, but utterly meaningless for our purposes); primary papers that use the algorithm (contribute in a sense to notability, but suffer from being primary); review articles, books etc. that discuss the algorithm in its context (secondary, the strongest evidence of notability, and the most useful to include as references). I would expect that a truly notable algorithm would eventually achieve some references in the third category; before it does, it might be WP:TOOSOON. I haven't expressed an opinion about this algorithm yet. I hope that clarifies a bit? Elemimele (talk) 16:49, 12 April 2022 (UTC)[reply]

Comment: I don't have a problem with Wikipedia documenting junk topics provided that we are generally able to document their unwholesomeness. But if the secondary literature fails to cast light the uselessness of what Eppstein's source calls useless metaphors, then we are better off without articles on them. The WP:SIGCOV criteria for the existence of sufficient sources to reach GNG requires that these sources be reliable, which means that they need to have 'editorial integrity'. In the context of peer-reviewed literature, we should demand that authors bear the burden of this editorial integrity, by critically evaluating the claims that they depend on in the literature that they cite. If work on 'metaphor-based metaheuristics' has secondary sources that critically evaluate the claim that the metaphors are not useless, then that reaches this bar for me, even if the secondary source says that the metaphor is in fact useless: if it amuses us, we can consider such work to be recreational computer science. However, I'll happily discount a dozen papers in high-impact venues that blindly build on the claims of an article if not one of them critically evaluates the significance claims. We cannot write neutrally on the topic if we lack sources with integrity. — Charles Stewart (talk) 09:47, 16 April 2022 (UTC)[reply]
Redirect to List of metaphor-based metaheuristics#A mayfly optimization algorithm. I spent a little time looking for sources that provide the kind of critical discussion I describe in my comment above of the core claims of the algorithm's proponents and found only one, The Challenge for the Nature-Inspired Global Optimization Algorithms: Non-Symmetric Benchmark Functions. While that article does have specific discussion of the article, and does generally talk about the problem of these kinds of paper cherry-picking the benchmarks that they present, this critical discussion I don't find has enough about this algorithm to write a balanced article. We're much less able to write a decent article on this topic, in view of the nature of the sourcing, than we are with the similar Firefly algorithm. Note that as an WP:ATD, a redirect outcome doesn't delete the page history, which would allow someone who found adequate sources to use the old page to write an expanded version, as GusRDRM proposed. I'd encourage anyone who wants to do this to check with other editors that the sources do suffice to properly treat the key claims made by the algorithm's proponents. I'll also not that List of metaphor-based metaheuristics has a serious problem with uncritically repeating the claims made in the abstract of the proposing papers: the target of the proposed redirect would benefit more from pruning than another merge. — Charles Stewart (talk) 13:10, 16 April 2022 (UTC)[reply]
Redirect to an appropriate section of List of metaphor-based metaheuristics as suggested above. I agree with that evaluation of the available sourcing. XOR'easter (talk) 01:14, 17 April 2022 (UTC)[reply]
I am focusing on gathering more sources that use the mayfly algorithm as the main algorithm, to improve the article. There may be a slight delay due to the holidays. Thank you! GusRDRM (talk) 10:59, 18 April 2022 (UTC)[reply]

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.