Stack Overflow does not present who voted for whom, but I wonder if it collects this data and make some intelligent deductions on it? For example, preventing a user from increasing his reputation through a fake account or a friend's account etc?
Besides fake account/voting detection, I wonder the way how smart deduction is done on any other data if there is any. For example click counts on user profiles or tags.
If there is some analytical mechanism behind, I wonder what kind of mechanism it is. Is there any rule engine or some conditions are check manually etc?
I found some articles about fake account/voting explaining the detection mechanism in an informal way like in the Anna Lear's answer. Is it possible to any find more formal explanation like is there an ontology being used or rules are hardcoded in the implementation.
I never worked on such a big project, so my question may be silly. Sorry for that.