Anti-spam and irrelevant content algorithm

Question

We are writing a service . I would like some advice on how to deal with spam and irrelevant responses on these sites. User data (a graph of connections between users and questions/answers with edge types: spam, positive answer, unsatisfactory answer, etc.)
There is an option to allow only people with close ranks (good bad answer ratio) to answer. Thus users with a high rating can only answer and ask users with the same high rating, but with a low one and beginners only with the same low one. There is a problem with newbies being more likely to receive low quality or relevant answers. What do you think?

Answer 1

I didn’t catch up a little, but why is the probabilistic model not suitable?
that is, a person with a high quality of answers has a greater “weight”, a person with a low one has a small one.
Well, randomly choose a group of, say, 10 people. Let me explain about “weight” - there is a person with a weight of 10, if with a weight of 1 and with a weight of 5, roughly speaking on a piece of paper we write down the name of a person with a weight of 10 10 times, 5 with a weight of 5 and 1 time with a weight of 1 and poke a finger. Then, with a high degree of probability, a high-quality answerer will come across, but there is still a chance to raise your rating for bad answerers.