Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Could anyone come up with a probability density function which is:

  • supported on [1,∞) (or [0,∞))
  • increasing
  • discrete
share|improve this question
    
you mean the density function? CDF is always increasing/non-decreasing. –  Seyhmus Güngören Mar 17 '13 at 15:47
    
@Seyhmus, yes . –  Yariv Mar 17 '13 at 15:48
1  
One cannot find such a density function: The integral cannot be $1$. –  André Nicolas Mar 17 '13 at 16:07

3 Answers 3

up vote 1 down vote accepted

There exist no such densities. Because, if a function is increasing, then either it has a limit and the function converges to this limit, or it doesnt have a limit.

If it doesnt have a limit then without any doubt the area under the function can not add up to $1$.

In case it is a convergent function and say it converges to $\alpha$, then an amount of area, say $\beta$ is repeated infinitely many times in the integration, therefore integral does not converge.

As a result, no such densities exist.

share|improve this answer
    
@Yariv Yes. Think about the uniform (discrete or continuous) distribution. It even does not increase but does not exists either. –  Seyhmus Güngören Mar 17 '13 at 16:18

A Dirac delta function at infinity.

share|improve this answer

I was having the same question, but in a bounded environment, so the distribution would be increasing with support a and b. It grows like a slow exponential, actually is the the distribution of the following:

Take N repetitions of 4 uniformly generated values between A and B. Make the histogram of the max in each repetition. You have something increasing that is not linear, it would be a kind of low speed growing exponential.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.