How to Generate random number that most of them ZEROs

Hi,

I would like to generate a set of 50 numbers that are in the range of [0-2] and follow uniformly distributed; however, I want most of them to be 0( Zeros).

I tried the following code but it does not generate most of them zeros.

N=randi([0 2],1,50)

Thank you

[Akira Agata]

"range of [0-2] and follow uniformly distributed" but "most of them to be 0" ? It's not clear for me what the 50 random number array you want looks like. Please tell us more details on what you would like to generate?

Answer 1

[Jeff Miller]

x=zeros(1,50);
r=rand(1,50);
PrZero = .8;  % Adjust as needed to achieve "most of them zeros"
x(r<(1-PrZero)/2) = 1;
x(r>1-(1-PrZero)/2) = 2;
histogram(x);

[zeezo]

Thank you very much. this is very helpful.

If I want to make a change to the range to be from 0 to 4 instead of 0 to 2, what should I do?

Answer 2

[dpb]

Well, if they're to be uniformly distributed then only 33% of them on average will be zero; that's what "uniform" means.

You can generate an array of N zeros and then distribute some M other values within it, but you won't be able to call it "uniformly distributed"

To answer the follow-up as well...

S=50;               % sample size
M=1; N=4;           % range other than zeros
x=zeros(S,1);       % initialize
k=round(S*(1-pz));  % pz ==> PrZero --> compute number nonzero
x(randperm(S,k))=randi([M N],1,k);   % spread that many around

[zeezo]

Thank you.

The code does not work it shows this error "Error using randperm K must be less than or equal to N."

[dpb]

Typo; I first had N as the sample size then changed it to S and missed the use in the argument to randperm

x(randperm(S,k))=...

made the correction in Answer...

Answer 3

[Jeff Miller]

It depends a little on whether you want (1) exactly the same number of zeros in each random set, or (2) a large number of zeros on average. I interpreted the question in the latter way (i.e., the number of zeros would fluctuate randomly). If that is indeed what you want, then it is probably easiest to do it like this:

S=50;               % sample size
M=1; N=4;           % range other than zeros
PrZero = .8;        % Whatever proportion of zeros you want on average.
x=randi([M N],1,S); % Uniformly distributed random numbers from M to N
r=rand(1,S);        % Random numbers from 0-1
x(r<PrZero) = 0;    % Select out random positions in x and overwrite them with zero
histogram(x)