search

jeffalstott / powerlaw

607 stars 134 forks source link

threshold in powerlaw fit #87

Closed silverbullet1472 closed 3 years ago

silverbullet1472 commented 3 years ago

Hi Jeff, Thanks you for powerlaw package. However, I noticed that in some literature, they fit power law with a "threshold", i.e., (r+r0) follows powerlaw. For example, see this Nature and PLOS One paper below. This PLOS One paper cited Clauset for distribution fitting. But it seems no option to find this threshold value in your package? So my question is, what is the difference between directly fit x and fit (x+threshold)? It seems the latter could provide a straight line in log-log plot, but lack practical meaning. I would be appreciated if you could offer your professional advice on this!~

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0037027 https://www.nature.com/articles/nature06958

jeffalstott commented 3 years ago

"It seems the latter could provide a straight line in log-log plot, but lack practical meaning." That's what I would say!

Thanks for using powerlaw!