# Assign Beta distribution to set of integers

1 view (last 30 days)
summyia qamar on 14 Jan 2018
Edited: John D'Errico on 14 Jan 2018
I have a column vector
x=[6.1000
9.4000
8.1000
3.2000
6.5000
7.2000
7.8000
4.9000
3.5000
6.6000
6.1000
5.1000
4.9000
4.2000
6.4000
8.1000
6.0000
8.2000
6.8000
5.9000
5.2000
6.5000
5.4000
5.9000
9.3000
5.4000
6.5000
7.4000
6.0000
12.6000
6.8000
5.6000
6.4000
9.5000
7.2000
5.6000
4.7000
4.5000
7.0000
7.7000
6.9000
5.4000
6.3000
8.1000
4.9000
5.3000
5.0000
4.7000
5.7000
4.9000
5.3000
6.4000
7.5000
4.4000
4.9000
7.6000
3.6000
8.3000
5.6000
6.2000
5.0000
7.4000
5.2000
5.0000
6.5000
8.0000
6.2000
5.0000
4.8000
6.2000
4.9000
7.0000
7.7000
4.7000
5.0000
6.0000
9.0000
5.7000
7.1000
5.0000
5.6000
4.9000
7.8000
7.1000
7.1000
11.5000
5.4000
5.2000
6.1000
6.8000
5.4000
3.5000]
I want to apply Beta distribution to the data but I cnt get any reasonable answer from matlab documentation. I tried fitdist but it didn't work. neither distributionfitter app help in these distributions. can anybody explain me this?
##### 2 CommentsShowHide 1 older comment
summyia qamar on 14 Jan 2018
I want to apply beta distribution. I have taken this example from a book'Simulation with arena' in which they used input analyzer and got the answer for square error for beta distribution. however Matlab say the values must be in interval [o,1]

John D'Errico on 14 Jan 2018
Edited: John D'Errico on 14 Jan 2018
These are not integers anyway.
But what stops you from normalizing them so they lie in [0,1]?
xmin = min(x);
xmax = max(x);
xhat = (x - xmin)/(xmax - xmin);
[PHAT, PCI] = betafit(xhat)
PHAT =
0.15292 0.39984
PCI =
0.14753 0.35227
0.1585 0.45383
So a beta distribution that lives on [xmin,xmax].