i should write a fuzzy finder library that gets a set of strings/objects and detects its repetition rate in the sample,
so if the content of the object is similar to the objects in the sample set, it returns a confidence rate of how close is it to the dataset samples, so there would be an easy way to detect spam by seeing it certain phrases are starting to get repeated and how many times
service that collects posts, and compares them to the rest of the posts to see if the content is similar to the other posts in the database.
if let's say, someone's spamming the same phrase, the service would notice that the sample rate for a certain set of strings is starting to build up, and flag them as spam
Add comment