Quote:
Originally Posted by brandonstills
The next best approach is some kind of heuristics based approach that involves machine learning. That is really hard to determine and implement. It is also very specific to your site and will need a lot of fine tuning. Another simple answer is to just rate limit the number of requests per IP.
|
Brandonstills thanks for your smart answer.
A heuristic approach is exactly what I had in mind.
An algorithm to which you can feed a site's average KPIs to be used as the control group. Ideally set them as a range MIN MAX.
Example (not exhaustive) Takes into account #of requests by IP (vs Human Request)
AND User Agent
AND BR% range
AND Page/Sess
AND TOS
AND USER FLOW (1st/ 2nd/ nth interaction compared to the MIN MAX users habits from the same traffic type)
Etc.
It is a good project. The one that makes it can sell it to many, many webmasters/ traffic brokers..