Прислал журнал статью на рецензию. Marine Studies. Почитал. Говно какоето.

Камрады. Если что-то для кого-то пишите - пишите нормально. Статья - 20 страниц воды, из которых 2/3 сгенерированные нейросетью. И статья про нервосеть.

Имейте уважение - продираться через такое - труд.

Ниже - то что ответил в редакцию.
Dear Jesnica!

Hope this mail finds you well!

I can not accept the invintation since my qualification is far from the topic. I am radioastronomer and my knowlege in marine science is very limited.

Still, after quick look. I can not recomend this article to review by anyone. The article is poorly written and the overcomplicated, but the results is limited. The paper contains 20 pages and can be shrinked in a half without any valuable scientific losses. The authors use five error measure indicators without argumentation why only one can not used. The authors do not provide any clarification why the ANN should be used in advance to other statistical methods. The authors do not provide enough information to conclude that model was not overfitted. The authors copmare the error measure indicators with other studies without clarification about the statistical importance of the indicators - the datasets is different and indicators can be biased by the datasets rather fitted model. In the present study, model was fitted with 87(!) samples. Such number is comparable to the number of the free parameters in ANN model and can not meet Nyqist criteria so the results obtined is poorly validated.

The paper is poorly written, has a lot of well-known propositions that looks like to be generated by ltsm models. The paper would be much more readable if authors focus on the scientific results if any.

Best, Petr Zemlyanukha