How to specific random numbers rang?

Hello.

This syntax gave me a specific mean (mu) and specific standard diviation but how to specific range of random numbers? such as between 1 to 9,like 5 2 3 4 1 5 6 7 7 9 but not 1.2 or 5.5

r=mu+sd.*randn(m,n);

all I need is a set of random numbers with the same mu, SD and range between 1 to 9.

Thankyou for your kineness and so sorry about my broken english ^^

Answer 1

[Paulo Silva]

randi([1 9],m,n)

Answer 2

[Matt Fig]

As Walter stated, this is not really possible exactly. However, you can get close (to some level of tolerance):

x = ceil(4.5 + .925.*randn(10000,1));  % The distribution
[min(x) max(x) mean(x)]  % look at range.
hist(x,unique(x));  % Look at distribution.

[Walter Roberson]

A quick test to see how many iterations you go before generating something that would be out of range:

T = 0;while true; T = T + 1;x = ceil(4.5 + 0.925*randn); if x < 1 || x > 9; disp([T,x]); break; end; end

Over 11 runs the average number of iterations in my trial was 670586, so as an approximation, one value would be out of range for every 2/3 of a million generated points.

[Matt Fig]

Good enough for government work...;-)

Answer 3

[Chanon]

r=mu+sd.*randn([1 9],m,n);

T-T

warning : input argument must be scalar

[Walter Roberson]

You did not use randi() like Paulo suggested.

Answer 4

[Walter Roberson]

It is not possible to constrain a normal distribution to a particular range. A normal distribution has an infinite tail in both directions. If you are required to select from a finite set of outcomes, then the distribution can never be a normal distribution (but it might be a good approximation if the number of different outcomes is large enough.)

Paulo's answer is fine if you want to select integers from 1 to 9 with equal probability (uniform distribution)

If you want to select 1 to 9, each with probabilities P(K) (i.e., P is a vector that gives the probabilities) then

[counts, bins] = histc(rand(1,n*m), [0 cumsum(P(1:end-1)) 1]; r = reshape(bins, n, m);

It is of course straight-forward to modify this to start at an integer other than 1.

Answer 5

[Chanon]

Million thanks for you guys, ^_^ I'm new for this MATLAB but for some reason I have to work with this also I'm an idiot for statistic too, this is so much help me solve the problem faster than reading the books, ^_^