How to generate a matrix with random numbers between 0 and 1 and mean of the distribution is unspecified.

Every time I generate matrix with random numbers between 0 and 1, mean of the generated distribution is different.

[Walter Roberson]

Are you asking how to deliberately generate a different mean each time? If so then I suggest generate random coefficients for a beta distribution.

Answer 1

[Torsten]

If you choose the number of random numbers large enough, their mean should approach 0.5:

n = 20000;
y = rand(n,1);
ym = mean(y)

If you want the same numbers ( and thus the same mean ) in each run, use

rng('default')

Answer 2

[John D'Errico]

You need to understand the difference between a sample mean, and the population mean.

x = rand(1,100);
mean(x)

ans = 0.4463

This is a random SAMPLE, so at best, it will only have a mean near the population mean for that uniform random number generator, which is theoretically 0.5 exactly.

With larger sample size, the sample mean approaches the population mean, but that is only an asymptotic thing.

mean(rand(1,1e6))

ans = 0.4999

If you really want a sample with the sample mean EXACTLY the same as the population mean, then your sample is not as easy to generate. However, it can be done using Roger Stafford's randfixedsum, as found on the file exchange. Better here is probably to just understand the difference.