generating random variable between 0 and 1 with probabilities

Hi all,

I have two questions please:

1) how do i generate a random number either 0 or 1 with equal chances of accept and reject 50-50 chance?

2) how do i generate a random number either 0 or 1 with percent shift in probablity in order to accept "bias' between -0.5 and 0.5

if percent shift=0 then 50-50 chance of accept

if percent shift=0.1 then 60% accept - 40% reject

if percent shift=0.2 then 70% accept - 30% reject

if percent shift=0.3 then 80% accept - 20% reject

if percent shift=0.4 then 90% accept - 20% reject

if percent shift=0.5 then 100% accept - 0% reject

if percent shift=-0.1 then 40% accept - 60% reject

if percent shift=-0.2 then 30% accept - 70% reject

if percent shift=-0.3 then 20% accept - 80% reject

if percent shiftt=-0.4 then 10% accept - 90% reject

if percent shift=-0.5 then 0% accept - 100% reject

Thanks!

[Walter Roberson]

If the percent_shift is 0.4 then you total 110% probability.

[william Smith]

sorry typo erro

Answer 1

[Walter Roberson]

 rand() < (percent_shift + 0.5)

[william Smith]

%Thank you!
%will this be the way to put it in a function:

i=1
function [choice_1] = customer_bias(customer_1,bias)
if bias(i)=0 %then 50-50 chance of accept
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=0.1  %60 accept - 40% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=0.2  %70% accept - 30% reject
   rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=0.3 then %80% accept - 20% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=0.4 then %90% accept - 10% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=0.5 then %100% accept - 0% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif percent shift=-0.1 then %40% accept - 60% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=-0.2 then %30% accept - 70% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=-0.3 then %20% accept - 80% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=-0.4 then %10% accept - 90% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=-0.5 then %0% accept - 100% reject
    rand(choice_1) < (bias(i) + 0.5)
i=i+1
end

%Thanks again

[Walter Roberson]

In MATLAB, = is only for assignment, not for comparison.

Also, what if the user passes something that is not bit-for-bit identical to how MATLAB convers the literal 0.3 (for example) ? Suppose they pass in 0.25 for the bias?

[william Smith]

%would this make a better sense?
i=1
function [choice_1] = customer_bias(customer_1,bias)
if bias(l)>=0 && bias(l)<0.1 %then 50-50 chance of accept
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(l)>= 0.1 && bias(l)<0.2  %60 accept - 40% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(l)>=0.2 && bias(l)<0.3  %70% accept - 30% reject
   rand(choice_1) < (bias(i) + 0.5)
elseif bias(l)>=0.3 && bias(l)<0.4 then %80% accept - 20% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(l)>=0.4 && bias(l)<0.5  %90% accept - 10% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=0.5 then %100% accept - 0% reject
    choice_=1;
elseif bias(l)>=-0.1 && bias(l)<-0.2 %40% accept - 60% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(l)>=-0.2 && bias(l)<-0.3  %30% accept - 70% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(l)>=-0.3 && bias(l)<-0.4  %20% accept - 80% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(l)>=-0.4 && bias(l)<-0.5%10% accept - 90% reject
    rand(choice_1) < (bias(i) + 0.5)
elseif bias(i)=-0.5 then %0% accept - 100% reject
    choice_1=0
i=i+1
end

[Walter Roberson]

The assignment to i before the function definition is in a different workspace than the function uses. Inside your function, both i (lower-case I) and l (lower-case L) are undefined. You also switch back and forth between using i or l as the subscript.

You assign a value to choice_ in the case of bias 0.5 but you never use that variable.

You only assign a value to the output variable choice_1 in the case that the input bias is -0.5

You try to test 0.5 and -0.5 for equality using = as the comparison operator, but = is only ever assignment (except some cases having to do with older versions of the symbolic toolbox.) == is the comparison test.

You input customer_1 but you never use it.

You use choice_1 to pass a size argument to rand() but you have not defined choice_1 by that point.

It is not clear what your intended output is. Is it the random variable such as 0.013433807 ? Is it 0.0 or 0.1 (double precision numbers) ? Is it uint8(0) or uint8(1) ? Is it false or true (logical values) ?

Why are you indexing bias? Are you expecting it to be a vector of values?

Each of your cases appears to have the same action except for 0.5 and -0.5 . Why not just test those two specifically, like

if bias(i) == 0.5
    choice_ = 1;
elseif bias(i) == -0.5
    choice_1 = 0;
else
    rand(choice_1) < (bias(i) + 0.5)
end

[william Smith]

Awsesome!

Thank you so much!!