Say I have a website with a lot of pages and traffic. Some of them are more visited, some of them a less. My task is to split pages into two groups for the given time period, so that the total traffic going to these groups is as similar as possible (day-wise). Could you please suggest a powerful method which could help me with this problem?

Thank You!

  • $\begingroup$ Sure: flip a coin. The computer can do that efficiently and repeatedly. $\endgroup$
    – whuber
    Sep 15, 2020 at 15:13
  • $\begingroup$ @whuber thank you for reply. Ok, I flip a coin N times (number of pages) and end up with two groups with almost the same number of pages, but with completely different traffic. That is not what I want. What I want is two groups with possibly different number of pages, but with the really close traffic behaviour. Say I have a daily traffic for last 30 days for every single page. I want to end up with two time series which are as closest as possible to each other (without lags considering, only daily difference). $\endgroup$
    – Thomas B.
    Sep 15, 2020 at 15:28
  • $\begingroup$ The randomness guarantees, with astronomically high probability, that the "traffic behavior," however you might define that, will be comparable in the two groups. $\endgroup$
    – whuber
    Sep 15, 2020 at 15:44


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.