I happened to encounter lots of scientific/business scenarios where a Zipf/Pareto/powerlaw describes well my data. However, whenever the mean of the distribution is large enough, the fact that these distributions have infinite variance turns into unphysical results. I then found articles mentioning "hooked powerlaw" that I guess is meant to address this. What's a mathematical formulation of it? Any python/numpy code for that?

  • $\begingroup$ Google takes you here: sourceforge.net/projects/hooked-power-law-r. $\endgroup$ – whuber Jul 27 '17 at 15:30
  • $\begingroup$ @whuber thanks but that is R code - guess I could try to understand it and post the math here/convert it to python. $\endgroup$ – famargar Jul 28 '17 at 9:15

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.