Triangular Distribution MoM Fit Details

Property 1: If you know the value of a then you can estimate b and c by

Estimating c from a

b = 3m – a – c

Proof: As shown in Method of Moments: Triangular Distribution, the mean m and variance v of a triangular distribution follow the following equations

m = (a + b + c)/3

v = (a2 + b2 + c2 – ab – ac – bc)/18

From the first equation, it follows that

b = 3m – a – c

b2 = (3m – a – c)2 = 9m2 + a2 + c2 6ma – 6mc + 2ac

From the equation for the variance, it follows that

18v = a2 + b2 + c2 – ab – ac – bc

Substituting the expressions for b and b2, it follows that

18v = a2 + 9m2 + a2 + c26ma – 6mc + 2ac + c2 – ab – ac – bc + c23ma + a2 + ac – ac – 3mc + ac + c2

= 3a2 + 9m2 +3c2 9ma 9mc + 3ac

This results in one equation in one unknown

6v = a2 + 3m2 +c2 3ma 3mc + ac

which is equivalent to

c2 + (a – 3m)c + (a2 + 3m23ma – 6v) = 0

Using the quadratic formula, it follows that

c formula 1

Actually, we can use the version with the positive sign before the square root symbol.

Formula for c

This has a solution provided 8v ≥ (ma)2.

Observation: Note that

b = 3m – a – c

and so b can also be calculated via the formula for c above except that the negative sign is used before the square root symbol.

References

Cook, J. (2015) Fitting a triangular distribution
https://www.johndcook.com/blog/2015/03/24/fitting-a-triangular-distribution/#:~:text=One%20way%20to%20fit%20a,if%20these%20values%20are%20known.

Kotz, S., van Dorp, R.(2004) The triangular distribution. Beyond Beta
https://books.google.it/books/about/Beyond_Beta_Other_Continuous_Families_Of.html?id=JO7ICgAAQBAJ&redir_esc=y

Leave a Comment