Proving Distribution Equality Using Stirling Approximation

  • Thread starter Thread starter physics.alex
  • Start date Start date
  • Tags Tags
    Distribution
Click For Summary

Homework Help Overview

The discussion revolves around proving a distribution equality using the classic Binomial distribution and Stirling approximation. The original poster presents a complex expression for the distribution and seeks guidance on simplifying the proof.

Discussion Character

  • Exploratory, Assumption checking, Mathematical reasoning

Approaches and Questions Raised

  • Participants suggest substituting the expression for I(x) into P(X) to simplify the proof. There is a focus on clarifying the ambiguous parts of the logarithmic expression, particularly regarding the correct interpretation of the terms involved.

Discussion Status

The discussion is ongoing, with participants providing suggestions for clarifying expressions and exploring the properties of logarithms. Some guidance has been offered regarding the use of Stirling approximations, but no consensus or complete solution has been reached.

Contextual Notes

There is an emphasis on the importance of clear notation, especially in the context of LaTeX formatting, to avoid ambiguity in mathematical expressions. The original poster expresses uncertainty about the notation used.

physics.alex
Messages
28
Reaction score
0

Homework Statement


I am asked to prove the following distribution equal to 1. The distribution is obtained by using classic Binomial distribution and apply stirling approximation.


Homework Equations



P(X) = \frac{1}{\sqrt{Nx(1-x)}}exp(-NI(x))

where I(x) = x ln (x/p) + (1-x) ln (1-x/1-p)


This form seems complicated to me. Any suggest for the first step to simplify the proof??
 
Physics news on Phys.org
I would say substitute the expression you have for I(x) into the expression you have for P(X), but what you have for I(x) is ambiguous, especially this part:
ln(1 - x/1 - p)

If I take this at face value, it is
ln(1 - \frac{x}{1} - p)

I don't think that's what you meant, though, so I will have to interpret what you have written.
Is it
ln(\frac{1 - x}{1 - p})?

Or is it
ln(1 - \frac{x}{1 - p})?
 
sorry for unclear formula.
it should be

ln{ (1-x) / (1-p) }

*I am not familiar with the latex format sorry..
Alex

Mark44 said:
I would say substitute the expression you have for I(x) into the expression you have for P(X), but what you have for I(x) is ambiguous, especially this part:
ln(1 - x/1 - p)

If I take this at face value, it is
ln(1 - \frac{x}{1} - p)

I don't think that's what you meant, though, so I will have to interpret what you have written.
Is it
ln(\frac{1 - x}{1 - p})?

Or is it
ln(1 - \frac{x}{1 - p})?
 
Whether you know LaTeX or not you should be aware of how to write rational expressions so that they are not ambiguous; that is, by using enough parentheses in the right places so that their meaning is clear.

For your problem, work with I(x) using the properties of logs. From this, you should get the log of the product of two expessions involving the fractions x/p and (1 - x)/(1 - p).

From there, replace the two factors by their Stirling approximations.

I haven't worked this all the way through, but that's the direction I would take.
 
Thanks and will try it.
Alex
 

Similar threads

  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 14 ·
Replies
14
Views
2K
  • · Replies 8 ·
Replies
8
Views
3K
Replies
4
Views
2K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 105 ·
4
Replies
105
Views
11K
Replies
1
Views
1K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 13 ·
Replies
13
Views
2K
  • · Replies 4 ·
Replies
4
Views
2K