1
$\begingroup$

Please look at the the image of the cdf and how it rotates about a specific point:

https://pubs.aeaweb.org/na101/home/literatum/publisher/aea/journals/content/aer/2006/aer.2006.96.issue-3/aer.96.3.756/production/pdfimages_v02/master.img-000.jpg

The tricky thing is to figure out how the parameters of the distribution map to the rotation point. ANY parameter change will cause a rotation somewhere, but I want to be able to control where the rotation occurs.

I think that the underlying pdf probably can't be symmetric if you want to control the rotation point. For instance, in the case of a normal distribution, you probably have to change both the mean and the variance in a particular way to control the rotation point.

How do I control the rotation point?

I would like to use a t distribution or normal distribution. I need to control the rotation point of a pdf as well.

$\endgroup$
2
  • $\begingroup$ You need to figure out the math that allows you to manipulate the CDF of a distribution the way you want. That is not a Mathematica problem. Once you have the math / statistics figured out, then attempt to implement the solution in Mathematica. If you get stuck anywhere along that path, we will be happy to help. $\endgroup$
    – MarcoB
    Jun 8 '18 at 4:37
  • $\begingroup$ @Mo_hazard Please do not vandalize your own question. $\endgroup$
    – halirutan
    Jun 11 '18 at 18:36
1
$\begingroup$

Maybe something like this: Using standard normal distribution (NormalDistribution[]) as the base, find the parameter μ[σ, x0] so that CDF of NormalDistribution[μ[σ, x0], σ] crosses the CDF of standard normal at x0:

ClearAll[x, x0, σ, mu]
mu[σ_, x0_] := First@FullSimplify[μ /. 
  Solve[{CDF[NormalDistribution[], x0] == 
     CDF[NormalDistribution[μ, Rationalize @ σ], Rationalize @ x0], 
     x0 < μ, σ < 1}, {μ}, Reals]]

Manipulate[dist = NormalDistribution[mu[σ, x0], σ]; 
    Labeled[Row[{Plot[Evaluate@{PDF[NormalDistribution[], x], PDF[dist, x]}, {x, -4, 4}, 
     Axes -> False, Frame -> True, 
     MeshFunctions -> {PDF[NormalDistribution[], #] - PDF[dist, #] &},
     Mesh -> {{0}}, MeshStyle -> PointSize[Large], ImageSize -> 300],
  Plot[Evaluate@{CDF[NormalDistribution[], x], CDF[dist, x]}, {x, -4, 4}, 
     Axes -> False, Frame -> True, 
     GridLines -> {{x0}, None}, ImageSize -> 300]}], 
  Style[#, 20] &@Row[{"x0 = ", x0, " && σ =", σ, " && mu[σ,x0] =", mu[σ, x0]}], Top],
 {{x0, -1}, -4, 0}, {{σ, .5}, .01, 1}]

enter image description here

enter image description here

$\endgroup$
2
  • $\begingroup$ " Solve was unable to solve the system with inexact coefficients. The answer was obtained by solving a corresponding exact system and numericizing the result." -- I am getting this error message. $\endgroup$
    – H_hazard
    Jun 11 '18 at 14:46
  • $\begingroup$ @Mo_hazard, I updated with a fix. Please let me know if the new version works. $\endgroup$
    – kglr
    Jun 12 '18 at 9:13

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.