0

Positively Skewed "Normal" Distribution

How to draw a positively skewed "normal" distribution, as shown above? I understand it's not even normal.

Ho Man-Ho
  • 171
  • You could warp the x axis (parametric). x=sqrt(t+a) might do,. x=log{t) (log normal) has historical value. – John Kormylo Sep 13 '20 at 15:57
  • @JohnKormylo Thx, but it doesn't seem to work. I tried $3*1/exp(((sqrt(\x+5)-3)^2)/2)$, but shows almost a flat curve. Domain is 0:6. Or did I do something wrong? – Ho Man-Ho Sep 13 '20 at 16:31
  • Ok, I got it. Simply multiply the error function (see https://en.wikipedia.org/wiki/Skew_normal_distribution) – Ho Man-Ho Sep 13 '20 at 17:07
  • 1
    Oops, that should have been x=exp(t) and, or y=exp(-t*t) etc. BTW, never use t^2 instead of t*t when computing. – John Kormylo Sep 13 '20 at 21:12

0 Answers0