How can a new user avoid having a legitimate mathematics question classified as spam?

I see the following problems with your original post.

(1) Your post included a link to Zenodo, which is probably what looked like spam. Typically, a question which requires the readers to look at another site is not suitable for MO. In this case, the link is to a 132-page document; it is unclear what is relevant, and if any of it is relevant, it should be reproduced. If reproducing the relevant portion makes your post very long, your post may just not be a good fit for MO.

(2) It is not true that "a well-defined mathematical question" is generally on-topic at MathOverflow, which is primarily for questions that professional mathematicians encounter in their research. This is often judged by the "level" of the mathematics in the post, though I wouldn't try to define level (and "low-level" questions that mathematicians find entertaining often do pretty well). I am not going to make a value judgement on the math in your post, but at first glance, I wouldn't be surprised if users viewed it as belonging more to Math StackExchange.

(3) Your post, and this meta post, have tics that lead me to believe they are AI output. Because the Zenodo link is to a document in German, I will assume these are artifacts of machine translation, and not that your entire post is AI-generated. I am not typically willing to make this assumption, and I suspect this is also the case for some other users who may have flagged your post as spam.

I suggested perhaps posting on Math StackExchange. I haven't used MSE in some years, but I am led to believe they are less tolerant of AI output than MO is. If you are able to write in English, you may find more success in writing your post without AI assistance. If not, I would still suggest to minimize AI use in writing your post; perhaps write the entire thing in German yourself, and then either run through Google Translate (which does not have the same tics) or ask an AI model to provide a translation which hews very closely to your original writing. If you do this, I would also suggest to state that you used AI for this purpose; if it is suspected that you did but you don't state that you did or for what, users may react more negatively to your post.