Difference between revisions of "Three parameter beta distribution estimation procedure"

From Open Food Facts wiki
Jump to navigation Jump to search
 
(3 intermediate revisions by the same user not shown)
Line 1: Line 1:
 +
Part of [[Nutritional values probability theory|this]].
 
No procedure for a three parameter estimation can be found on Internet, so we have to come up with something.
 
No procedure for a three parameter estimation can be found on Internet, so we have to come up with something.
  
Line 4: Line 5:
 
[[File:BetaDistributionFourParameters.png|300 px|thumb|right|Form. 1.]]
 
[[File:BetaDistributionFourParameters.png|300 px|thumb|right|Form. 1.]]
 
This formula has two extra parameters ''a'' and ''c'', which can be used to normalise the distribution to [0,1] with ''y = (c-a)x + a''. As nutritional values always start at 0, we assume that ''a = 0''. With this we can rewrite formula 1:
 
This formula has two extra parameters ''a'' and ''c'', which can be used to normalise the distribution to [0,1] with ''y = (c-a)x + a''. As nutritional values always start at 0, we assume that ''a = 0''. With this we can rewrite formula 1:
 +
 +
As MediaWiki does not supports maths (yet), have a look at [https://www.mathcha.io/editor/pPzgOF6yI3nHLmp8K9COol96JcN2Ypnos2Ln7e this]
 +
 +
A first experiment with estimating yielded no results: I did not manage to find the variables that converged to a solution. To be continued.

Latest revision as of 07:31, 11 October 2024

Part of this. No procedure for a three parameter estimation can be found on Internet, so we have to come up with something.

The starting point is the four parameter Beta distribution found on wikipedia:

Form. 1.

This formula has two extra parameters a and c, which can be used to normalise the distribution to [0,1] with y = (c-a)x + a. As nutritional values always start at 0, we assume that a = 0. With this we can rewrite formula 1:

As MediaWiki does not supports maths (yet), have a look at this

A first experiment with estimating yielded no results: I did not manage to find the variables that converged to a solution. To be continued.