How to generate samples from a modified exponential density?

Hi everyone, The standard form of an exponential density is f(x|lambda)= lambda*exp(-x/lambda) if x>=0 and f(x|lambda)=0 if x<0.

I want to generate samples from the following density f(x|lambda, b)= lambda*exp(-(x-b)/lambda) if x>b and f(x|lambda,b)=0 if x<=b. Where b is a user-specified parameter.

Can anyone here suggest me how to do this ?

Best regards,

Answer 1

[Star Strider]

This works:

lambda = 0.1;
m = 1;
n = 100;

expshft = @(b,lambda,x) (x>=b) .* lambda.*exp(-max(0, (x-b))./(b+lambda));

x = 25*rand(m,n);
R2 = expshft(10,lambda,x);

figure(1)
plot(x, R2,'*b')
grid

The actual function is ‘expshift’. The rest is demonstration code to show how it works.

[Cuong]

Thank you so much,

Your answer is very useful.

Best,

[Star Strider]

My pleasure!

That was an interesting problem!

Answer 2

[Shashank Prasanna]

You can generate uniform random numbers using the rand function and apply it to your modified density function.

You can create functions this way: http://www.mathworks.com/help/matlab/ref/function.html