Prove that the function given with f(x) is a PDF

  • Thread starter Thread starter MrKushtrim
  • Start date Start date
  • Tags Tags
    Function Pdf
Click For Summary

Homework Help Overview

The discussion revolves around proving that a given function, specifically f(x) = \binom{x-1}{r-1} p^{r}(1-p)^{x-r}, qualifies as a probability density function (PDF) for x = r, r+1,..., ∞ and 0 < p < 1. Participants explore the properties and requirements for this function to be considered a legitimate PDF.

Discussion Character

  • Exploratory, Assumption checking, Mathematical reasoning

Approaches and Questions Raised

  • Participants discuss the necessity of demonstrating that the sum of the function over its domain equals 1. Some suggest using the binomial theorem and series expansions, while others question the positivity of the function and the implications of using different variable notations.

Discussion Status

There is an ongoing exploration of the mathematical properties required for the function to be a probability mass function, with some participants providing guidance on the necessary conditions. Multiple interpretations of the function's properties are being examined, particularly in relation to series summation and binomial expansions.

Contextual Notes

Participants note the distinction between probability density functions and probability mass functions, emphasizing the need for clarity in terminology and the conditions under which the function is defined. There is also mention of constraints related to the parameters of the function, such as the range of p.

MrKushtrim
Messages
3
Reaction score
0
Hi,
Could someone help me prove that the function given with
##f(x) = \binom{x-1}{r-1} p^{r}(1-p)^{x-r}## is a probability density function, where ## x= r, r+1,..., \infty ## and ## 0<p<1 ##


I thought to solve it somehow by using the binomial theorem, but since it's the upper part that's changing on the binomial coefficients, this proved futile.
 
Last edited:
Physics news on Phys.org
Let's start simple:

Can you prove that the value is always larger than zero?
 
Use ## instead of $ for inline LaTeX.
 
MrKushtrim said:
Hi,
Could someone help me prove that the function given with
[tex]f(x) = \binom{x-1}{r-1} p^{r}(1-p)^{x-r}[/tex] is a probability density function, where [itex]x= r, r+1,..., \infty[/itex] and [itex]0<p<1[/itex]


I thought to solve it somehow by using the binomial theorem, but since it's the upper part that's changing on the binomial coefficients, this proved futile.

You need to show that [tex]\sum_{x=r}^\infty \binom{x-1}{r-1} p^r (1-p)^{x-r} = 1.[/tex] Setting [itex]q = 1- p[/itex] and [itex]m = x - r[/itex] gives [tex] \sum_{x=r}^\infty \binom{x-1}{r-1} p^r (1-p)^{x-r} <br /> = (1 - q)^r \sum_{m=0}^\infty \frac{(m + r - 1)!}{(r-1)!m!}q^m.[/tex] What does the series on the right have to equal if the left hand side is to equal 1, and can you prove that it does?
 
MrKushtrim said:
Hi,
Could someone help me prove that the function given with
$f(x) = \binom{x-1}{r-1} p^{r}(1-p)^{x-r}$ is a probability density function, where $x= r, r+1,..., \infty $ and $0<p<1$


I thought to solve it somehow by using the binomial theorem, but since it's the upper part that's changing on the binomial coefficients, this proved futile.

For discrete random variables, I prefer to use symbols like ##j## and ##k## instead of ##x##. So, in this notation: what properties must be possessed by
[tex]f_k = {k-1 \choose r-1} p^r (1-p)^{k-r}, \, k = r, r+1, r+2, \ldots[/tex]
in order that it be a legitimate probability mass function of a discrete random variable? Note that ##f_k## is NOT a probability density function; it is a probability mass function. There is a world of difference!
 
Hi, thanks for the reply .
pasmith said:
You need to show that [tex]\sum_{x=r}^\infty \binom{x-1}{r-1} p^r (1-p)^{x-r} = 1.[/tex] Setting [itex]q = 1- p[/itex] and [itex]m = x - r[/itex] gives [tex] \sum_{x=r}^\infty \binom{x-1}{r-1} p^r (1-p)^{x-r} <br /> = (1 - q)^r \sum_{m=0}^\infty \frac{(m + r - 1)!}{(r-1)!m!}q^m.[/tex] What does the series on the right have to equal if the left hand side is to equal 1, and can you prove that it does?
It must equal 1. Though I have no idea how to prove it.
Ray Vickson said:
For discrete random variables, I prefer to use symbols like ##j## and ##k## instead of ##x##. So, in this notation: what properties must be possessed by
[tex]f_k = {k-1 \choose r-1} p^r (1-p)^{k-r}, \, k = r, r+1, r+2, \ldots[/tex]
in order that it be a legitimate probability mass function of a discrete random variable? Note that ##f_k## is NOT a probability density function; it is a probability mass function. There is a world of difference!
Thanks, I had the terminology mixed up.
To be a probability mass function , ##f_k## has to be positive, and the sum over all ## k ## has to be 1.
It's easy to see that ##f_k## is positive, but I cannot seem to prove the second property
 
MrKushtrim said:
It must equal 1. Though I have no idea how to prove it.

Remember, you need to show that
[tex] (1 -q)^r \sum_{m=0}^\infty \frac{(m + r - 1)!}{(r-1)!m!}q^m = 1[/tex] for every [itex]0 < q < 1[/itex]. This suggests that the series [tex]\sum_{m=0}^\infty \frac{(m + r - 1)!}{(r-1)!m!}q^m[/tex] must be the binomial expansion of ... what function of [itex]q[/itex]?
 
  • Like
Likes   Reactions: 1 person
pasmith said:
Remember, you need to show that
[tex] (1 -q)^r \sum_{m=0}^\infty \frac{(m + r - 1)!}{(r-1)!m!}q^m = 1[/tex] for every [itex]0 < q < 1[/itex]. This suggests that the series [tex]\sum_{m=0}^\infty \frac{(m + r - 1)!}{(r-1)!m!}q^m[/tex] must be the binomial expansion of ... what function of [itex]q[/itex]?

## \frac{1}{(1 -q)^r } ##
But how would I prove that ?
---
Edit:
I checked a wikipedia article about binomial series, and it turns out that the series [tex]\sum_{m=0}^\infty \frac{(m + r - 1)!}{(r-1)!m!}q^m[/tex] is of this form , where ##β= r-1## and ##z=q##

Thanks for the help :)
Though, is there any other way to prove it, without directly using the formula ?
 
Last edited:
MrKushtrim said:
## \frac{1}{(1 -q)^r } ##
But how would I prove that ?
---
Edit:
I checked a wikipedia article about binomial series, and it turns out that the series [tex]\sum_{m=0}^\infty \frac{(m + r - 1)!}{(r-1)!m!}q^m[/tex] is of this form , where ##β= r-1## and ##z=q##

Thanks for the help :)
Though, is there any other way to prove it, without directly using the formula ?

Ultimately the only way to sum power series is to relate them to power series of known functions.

There are other ways of showing that [itex]\sum_{m=0}^\infty \frac{(m + r - 1)!}{(r-1)!m!}q^m = (1 - q)^{-r}[/itex]. For example, using the formula [tex] \frac{d^k}{dx^k} \left( \sum_{n=0}^\infty a_n x^n \right) <br /> = \sum_{n=0}^\infty a_{n+k} \frac{(n + k)!}{n!} x^n[/tex] we see that [tex] \sum_{m=0}^\infty \frac{(m + r - 1)!}{(r-1)!m!}q^m = \frac{1}{(r - 1)!} \frac{d^{r-1}}{dq^{r-1}} \left( \sum_{m=0}^\infty q^m \right).[/tex] [itex]\sum_{m=0}^\infty q^m[/itex] is a geometric series and is easily summed.
 

Similar threads

  • · Replies 14 ·
Replies
14
Views
3K
  • · Replies 13 ·
Replies
13
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
Replies
7
Views
2K
  • · Replies 20 ·
Replies
20
Views
3K
  • · Replies 19 ·
Replies
19
Views
4K
Replies
2
Views
1K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
3
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K