Distribution until the First Increase?

In summary: I realized E[N] was what was being asked, and I figured it was a matter of changing the sum to go from k=1 to infinity, because we were no longer starting at k=2. So it would be:\sum_{k=1}^{\infty} \frac{k}{k!}=k\sum_{k=1}^{\infty} \frac{1}{k!}=k(e-1)So the final result would be (e-1).I just wanted to correct my post from last night.
  • #1
Mbaboy
19
0

Homework Statement



Let [tex]X_1, X_2,...[/tex] be independent identically distributed continuous random variables. We call [tex]n[/tex] the first time of increase if

[tex]X_1>X_2>X_3>...>X_{n-1}<X_n[/tex]

Let [tex]N[/tex] be the time until the first increase. Show that [tex]E[N]=e[/tex].

Homework Equations



Given for a discrete random variable with values in the positive integers:

[tex]E(X)=\sum_{k=1}^\infty P(X>=k)[/tex]

The Attempt at a Solution



Well I know that [tex]X_i[/tex] are exchangeable so the desired outcome has only 1 in [tex]n![/tex] chance of happening.

[tex]P(N=n)=P(X_1>\cdots>X_{n-1}\mbox{ and }X_{n-1}<X_n)[/tex]

[tex]=P(X_1>X_2)P(X_2>X_3)P(X_3>X_4)...P(X_{n-2}>X_{n-1})P(X_{N-1}<Xn)[/tex]

[tex]=\frac{1}{2} \times \frac{1}{2^2} \times \frac{1}{2^3}...\frac{1}{2^{n-1}}[/tex]

Ok clearly I don't know what I'm doing. Maybe if someone could tell me what this evaluates to I could find my way to it? I just don't see how an [tex]e[/tex] can pop out whereas the only two ways to calculate [tex]e[/tex] I know are

[tex]\lim_{k\to\infty}(1+\frac{1}{k})^k[/tex] and [tex]e^z=\sum_{i=1}^\infty \frac{z^i}{i!}[/tex]

Also, I have no idea how the tail sum formula is suppose to help. How can I make use of it when the random variables in question are continuous?

Thanks guys!
 
Physics news on Phys.org
  • #2
These sort of questions never stop amazing me...

First, you got to keep in mind that although X is determined by a continuous random variable, the distribution of the number [tex]k[/tex] for which [tex]X_{k-1}<X_{k}[/tex] (+ the other condition) is, in fact, a discrete random variable. Let's call this discrete random variable [tex]Y[/tex] to avoid confusion with the other random variable [tex]X[/tex]. Second, the false assumption you have made so far is that you can factorize [tex]P(X_1>X_2>X_3)[/tex] into [tex]P(X_1>X_2)P(X_2>X_3)[/tex]. This is not true, since these are not independent outcomes. Namely, there are plenty of possibilites where [tex]X_2[/tex] can be larger than [tex]X_3[/tex], while at the same time be larger than [tex]X_1[/tex] - these are not independent.

You want to approach the situation a bit different.

First, note that [tex]Y[/tex] is in fact a discrete random variable, taking on the values [tex]2,3,4,\ldots[/tex]. The first thing you want to determine is: what is the distribution of [tex]Y[/tex]? In other words, what is

[tex]P(Y=k)[/tex]

For that we take the first [tex]k[/tex] outcomes [tex]X_i[/tex]. ([tex]i=1,\ldots,k[/tex]). We can always order these from highest to lowest:
[tex]X_{i_1} > X_{i_2} > \ldots > X_{i_k}[/tex]
for instance, we can have:
[tex]X_{4} > X_{8} > \ldots > X_2}[/tex]

There are in total [tex]k![/tex] ways of possible orderings. The important remark is now that all these orderings are equally likely to happen. And also, what we want is a particular subset of orderings, namely the one where we have
[tex]X_1 > X_2 > \ldots X_{k-1} < X_k[/tex]
You should convince yourself (again) that there are [tex]k-1[/tex] orderings corresponding to this statement. In general, the outcome [tex]X_k[/tex] is "plugged in" on the other side:

[tex]X_1> X_2 >\ldots X_i > X_k > X_{i+1} > \ldots X_{k-1}[/tex]

Again, there are [tex](k-1)[/tex] ways for this to occure. Since all possible orderings are equally likely to happen, the chance for the above to happen is:

[tex]P(Y=k) = \frac{k-1}{k!}[/tex]

Beautiful argument hey?

The rest of the question is now pretty straightforward, if you use the formula for the expectation value:

[tex]E[Y] = \sum{k=2}^\infty k P(Y=k)[/tex]
 
  • #3
Ah ok it all makes very good sense. The only thing I cannot grasp is how there are k-1 orderings. The way I see it there is only 1. Consider for k=4

[tex]X_1>X_2>X_3<X_4[/tex]

There is only 1 way to order it such that the condition still holds, or am I missing something?

Edit: Ok, I see for the random draws there are (k-1)! orderings so for

[tex]X_{(1)}>X_{(2)}>...>X_{(k-1)}[/tex]

but there is still only 1 ordering of the random variables. Maybe I am misinterpreting the question.

If you are considering the orderings and not the act
 
Last edited:
  • #4
xepma said:
These sort of questions never stop amazing me...

[tex]E[Y] = \sum{k=2}^\infty k P(Y=k)[/tex]

I presume that you suggest one ends up with;

[tex] \sum_{k=0}^{\infty} \frac{1}{k!} =e [/tex]

right?

But not E[Y] is asked but E[N] so I immediatley think of Walds equation (though I'm sure if it's applicable here but I'm pretty certain it is).
 
  • #5
Y is N.
 
  • #6
Mbaboy said:
Ah ok it all makes very good sense. The only thing I cannot grasp is how there are k-1 orderings. The way I see it there is only 1. Consider for k=4

[tex]X_1>X_2>X_3<X_4[/tex]

There is only 1 way to order it such that the condition still holds, or am I missing something?

Edit: Ok, I see for the random draws there are (k-1)! orderings so for

[tex]X_{(1)}>X_{(2)}>...>X_{(k-1)}[/tex]

but there is still only 1 ordering of the random variables. Maybe I am misinterpreting the question.

If you are considering the orderings and not the act

No, that is because this ordering:

[tex]X_1>X_2>X_3<X_4[/tex]

is not a unique one. Namely, you haven't specified the relation between [tex]X_4[/tex] and [tex]X_1[/tex] and [tex]X_2[/tex]. You can have:

[tex]X_1>X_2>X_4>X_3[/tex]
[tex]X_1>X_4>X_2>X_3[/tex]
[tex]X_4>X_1>X_2>X_3[/tex]

So there are 3 definite orderings corresponding to the one you want. The total number of orderings is [tex]4![/tex], so that

[tex]P(Y=4) = \frac{3}{4!}[/tex]

So again, in general you would have [tex]k![/tex] total orderings and [tex](k-1)[/tex] of them correspond to the case where
[tex]X_1>X_2>\ldots>X_{k-1} <X_k[/tex]
 
  • #7
dirk_mec1 said:
I presume that you suggest one ends up with;

[tex] \sum_{k=0}^{\infty} \frac{1}{k!} =e [/tex]

right?

But not E[Y] is asked but E[N] so I immediatley think of Walds equation (though I'm sure if it's applicable here but I'm pretty certain it is).

Yea, the forums were acting fishy last night, and it wouldn't let me edit the post.
 

FAQ: Distribution until the First Increase?

What is "Distribution until the First Increase"?

"Distribution until the First Increase" is a statistical concept that refers to the distribution of a variable until it reaches its first increase or improvement. It is commonly used in analyzing data related to economic growth, market trends, and other areas where changes over time are important.

How is "Distribution until the First Increase" calculated?

The calculation for "Distribution until the First Increase" involves identifying the point at which the variable starts to increase and then dividing the data into two parts: the values before the increase and the values after the increase. The distribution of the variable can then be analyzed separately for each part.

Why is "Distribution until the First Increase" important?

"Distribution until the First Increase" is important because it can provide valuable insights into the behavior and patterns of a variable over time. By separating the data into two parts, it allows for a more detailed analysis of how the variable changes before and after the first increase, which can be useful in decision making and forecasting.

Can "Distribution until the First Increase" be used for any type of data?

Yes, "Distribution until the First Increase" can be used for any type of data that shows a clear trend or pattern over time. This includes economic data, market data, and even scientific data such as climate change or population growth.

How can "Distribution until the First Increase" be visualized?

"Distribution until the First Increase" can be visualized in various ways, depending on the type of data and the purpose of the analysis. Common methods include line graphs, scatter plots, and bar charts, which can show the distribution of the variable before and after the first increase, as well as the overall trend over time.

Similar threads

Replies
1
Views
1K
Replies
9
Views
1K
Replies
3
Views
651
Replies
5
Views
2K
Replies
4
Views
613
Replies
1
Views
1K
Replies
8
Views
991
Back
Top