How can I make a custom random number distribution?

31 Ansichten (letzte 30 Tage)
Taneli Hautaniemi
Taneli Hautaniemi am 13 Jul. 2017
Beantwortet: Ka Mirul am 14 Nov. 2017
Hi! How can I use the inverse transform sampling in MATLAB to set the distributions of my random numbers? I know that normally distributed randn is just rand with a Gaussian distribution, so the latter should somehow be replaced with a custom pdf to generate a new function randc to be used in similar manner than the others. I thank for your help in advance!
  1 Kommentar
James Tursa
James Tursa am 13 Jul. 2017
What is your custom pdf based on? A formula description? A set of data points that you want to fit a pdf to? Or ...?

Melden Sie sich an, um zu kommentieren.

Antworten (3)

Image Analyst
Image Analyst am 13 Jul. 2017
You can do your own if you know the function and can get the CDF of the probability density function. You can then use inverse transform sampling. See Wikipedia: http://en.wikipedia.org/wiki/Inverse_transform_sampling
I attach an example for the Rayleigh distribution. Adapt as needed, for example to use an ad hoc, discrete distribution rather than an analytical one with a formula.

Star Strider
Star Strider am 13 Jul. 2017
In the Statistics and Machine Learning Toolbox, you can generate random numbers from any of the supported Continuous Distributions (link) or discrete or multivariate distributions (listed in the Contents panel on the left of the page).
  1 Kommentar
Taneli Hautaniemi
Taneli Hautaniemi am 13 Jul. 2017
Thank you but isn't what I'm looking for. I want to use my own distribution with my own function.

Melden Sie sich an, um zu kommentieren.


Ka Mirul
Ka Mirul am 14 Nov. 2017
I have create a video about generating random number in MATLAB
https://www.youtube.com/watch?v=MyH3-ONYL_k
Hope that will help you

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by