Prev: for loop imread
Next: Unexplainable Mex Error
From: Jon Smith on 9 Jun 2010 15:58
Does anyone know how to simulate a logit error? Thanks!
From: Tom Lane on 9 Jun 2010 17:48
> Does anyone know how to simulate a logit error? Thanks!

Jon, I am not sure this is what you are asking, but here is a script that
simulates values from a logistic distribution with location parameter 100
and scale parameter 10, and compares a histogram of the data to the
theoretical density.

If you're looking for something else, please elaborate.

-- Tom

z = random('logistic',100,10,[1000,1]);
hist(z,5:10:195)
area = length(z) * 10; % #points * binwidth
x = 0:200;
y = area * pdf('logistic',x,100,10);
line(x,y,'color','r','linewidth',2)


From: Jon Smith on 10 Jun 2010 09:23
Thanks Tom. That seems to be what I am looking for, but after the first line, Matlab gives me the error "Unrecognized distribution name: 'logistic'." Any thoughts?


"Tom Lane" <tlaneATmathworksDOTcom(a)nospam.com> wrote in message <hup27m$lh0$1(a)fred.mathworks.com>...
> > Does anyone know how to simulate a logit error? Thanks!
>
> Jon, I am not sure this is what you are asking, but here is a script that
> simulates values from a logistic distribution with location parameter 100
> and scale parameter 10, and compares a histogram of the data to the
> theoretical density.
>
> If you're looking for something else, please elaborate.
>
> -- Tom
>
> z = random('logistic',100,10,[1000,1]);
> hist(z,5:10:195)
> area = length(z) * 10; % #points * binwidth
> x = 0:200;
> y = area * pdf('logistic',x,100,10);
> line(x,y,'color','r','linewidth',2)
>
From: Tom Lane on 10 Jun 2010 10:36
> Thanks Tom. That seems to be what I am looking for, but after the first
> line, Matlab gives me the error "Unrecognized distribution name:
> 'logistic'." Any thoughts?
....
>> z = random('logistic',100,10,[1000,1]);

Jon, I would expect that to work in relatively recent versions of MATLAB. It
turns out the logistic distribution has a very simple form, though. It's
easy enough to generate random uniform numbers and invert the cdf. You can
replace this line by

p = rand(1000,1); z = 100 + 10*log((1-p)./p);

-- Tom


 | 
Pages: 1
Prev: for loop imread
Next: Unexplainable Mex Error