the beta distribution

logistic_guy

Senior Member
Joined
Apr 17, 2024
Messages
1,537
here is the question

Derive the mean and variance of the beta distribution.


my attemb
i want to show the mean = μ=αα+β\displaystyle \mu = \frac{\alpha}{\alpha + \beta}
i want to show the variance = σ2=αβ(α+β)2(α+β+1)\displaystyle \sigma^2 = \frac{\alpha \beta}{(\alpha + \beta)^2(\alpha + \beta + 1)}

the beta function B(α,β)=Γ(α)Γ(β)Γ(α+β)\displaystyle B(\alpha,\beta) = \frac{\Gamma(\alpha)\Gamma(\beta)}{\Gamma(\alpha + \beta)}
how this information help me driving the answer
 
here is the question

Derive the mean and variance of the beta distribution.


my attemb
i want to show the mean = μ=αα+β\displaystyle \mu = \frac{\alpha}{\alpha + \beta}
i want to show the variance = σ2=αβ(α+β)2(α+β+1)\displaystyle \sigma^2 = \frac{\alpha \beta}{(\alpha + \beta)^2(\alpha + \beta + 1)}

the beta function B(α,β)=Γ(α)Γ(β)Γ(α+β)\displaystyle B(\alpha,\beta) = \frac{\Gamma(\alpha)\Gamma(\beta)}{\Gamma(\alpha + \beta)}
how this information help me driving the answer
Very "poor" question!!

what is α ? what the definition of α?

what is ß ? what the definition of ß?
 
why?☹️


α\displaystyle \alpha and β\displaystyle \beta isn't part of the question
That's not true. The cumulative distribution function (CDF) and/or the density function is part of the question since it is necessary to compute mean and variance, and the parameters α,β \alpha,\beta are part of these functions. How do you define the beta-distribution?

Wikipedia has quite some information about it: https://en.wikipedia.org/wiki/Beta_distribution
 
I don't know what to tell you. The integration is a bit complicated but what's shown there is straightforward.

If you're at the level where you've been tasked to find the mean and variance of the Beta distribution you should be able to follow that integration.
 
thank

μ=01xxα1(x1)β1B(α,β)dx=αα+β\displaystyle \mu = \int_{0}^{1}x\frac{x^{\alpha - 1}(x - 1)^{\beta - 1}}{B(\alpha,\beta)}dx = \frac{\alpha}{\alpha + \beta}

it don't explain how he did this step☹️
If you followed what professor fresh_42 has given you in post #4, you would have already solved this integral by now. Instead, you have involved in an unnecessary conversation.

How to solve this integral?
The first step is to pull out any constants and to leave inside the integral only the things that have the variable xx. Do it, what do you get?
 
If you followed what professor fresh_42 has given you in post #4, you would have already solved this integral by now. Instead, you have involved in an unnecessary conversation.

How to solve this integral?
The first step is to pull out any constants and to leave inside the integral only the things that have the variable xx. Do it, what do you get?
thank

μ=1B(α,β)01xxα1(x1)β11dx\displaystyle \mu = \frac{1}{B(\alpha,\beta)}\int_{0}^{1}x\frac{x^{\alpha - 1}(x - 1)^{\beta - 1}}{1}dx

B(α,β)=Γ(α)Γ(β)Γ(α+β)\displaystyle B(\alpha,\beta) = \frac{\Gamma(\alpha)\Gamma(\beta)}{\Gamma(\alpha + \beta)} so it don't have x\displaystyle x
 
α, β\alpha,~\beta are clearly the parameters of the beta distribution.

Further explanation shouldn't be necessary.
I thought the point was that the OP understands why there are such parameters and their role in the distribution. Otherwise, he could have just asked for the integrals.
 
thank

μ=1B(α,β)01xxα1(x1)β11dx\displaystyle \mu = \frac{1}{B(\alpha,\beta)}\int_{0}^{1}x\frac{x^{\alpha - 1}(x - 1)^{\beta - 1}}{1}dx

B(α,β)=Γ(α)Γ(β)Γ(α+β)\displaystyle B(\alpha,\beta) = \frac{\Gamma(\alpha)\Gamma(\beta)}{\Gamma(\alpha + \beta)} so it don't have x\displaystyle x
Very good. And when you simplify it, you will get:

μ=1B(α,β)01xα(x1)β1 dx\displaystyle \mu = \frac{1}{B(\alpha,\beta)}\int_{0}^{1}x^{\alpha}(x - 1)^{\beta - 1} \ dx

I am sure that by mistake you wrote (1x)β1(1 - x)^{\beta - 1} as (x1)β1(x - 1)^{\beta - 1}, so the correct integral is:

μ=1B(α,β)01xα(1x)β1 dx\displaystyle \mu = \frac{1}{B(\alpha,\beta)}\int_{0}^{1}x^{\alpha}(1 - x)^{\beta - 1} \ dx


In the OP, you have told us that:

B(α,β)=Γ(α)Γ(β)Γ(α+β)\displaystyle B(\alpha,\beta) = \frac{\Gamma(\alpha)\Gamma(\beta)}{\Gamma(\alpha + \beta)}

Instead of that, you should have told us that:

B(α,β)=01xα1(1x)β1 dx=Γ(α)Γ(β)Γ(α+β)\displaystyle B(\alpha,\beta) = \int_{0}^{1}x^{\alpha - 1}(1 - x)^{\beta - 1} \ dx = \frac{\Gamma(\alpha)\Gamma(\beta)}{\Gamma(\alpha + \beta)}


Now compare between:

First integral:

01xα(1x)β1 dx\displaystyle \int_{0}^{1}x^{\alpha}(1 - x)^{\beta - 1} \ dx

And

Second integral:

01xα1(1x)β1 dx\displaystyle \int_{0}^{1}x^{\alpha - 1}(1 - x)^{\beta - 1} \ dx

What do you notice? And what should you do to make the first integral to be a beta function? Think a little and I am sure that you will see the trick (the idea)!

💡
 
What do you notice?
i notice the difference xα\displaystyle x^{\alpha} and xα1\displaystyle x^{\alpha - 1}

And what should you do to make the first integral to be a beta function?
i think i need -1 to make it beta function. i think it's wrong to just add -1
is the trick u\displaystyle u subtution?

u=xα\displaystyle u = x^{\alpha}
du=αxα1dx\displaystyle du = \alpha x^{\alpha - 1} dx

01xα(1x)β1dx=01u(1uα)β1duαxα1\displaystyle \int_{0}^{1}x^{\alpha}(1 - x)^{\beta - 1}dx = \int_{0}^{1}u(1 - \sqrt[\alpha]{u})^{\beta - 1}\frac{du}{\alpha x^{\alpha - 1}}

it become more complicated i don't see the trick☹️
i study integration long time ago i forgot all tricks
 
i notice the difference xα\displaystyle x^{\alpha} and xα1\displaystyle x^{\alpha - 1}


i think i need -1 to make it beta function. i think it's wrong to just add -1
is the trick u\displaystyle u subtution?

u=xα\displaystyle u = x^{\alpha}
du=αxα1dx\displaystyle du = \alpha x^{\alpha - 1} dx

01xα(1x)β1dx=01u(1uα)β1duαxα1\displaystyle \int_{0}^{1}x^{\alpha}(1 - x)^{\beta - 1}dx = \int_{0}^{1}u(1 - \sqrt[\alpha]{u})^{\beta - 1}\frac{du}{\alpha x^{\alpha - 1}}

it become more complicated i don't see the trick☹️
i study integration long time ago i forgot all tricks
Nice try, but that was not the trick. The trick is:

xα=xα+11x^{\alpha} = x^{\alpha + 1 - 1}

Let k=α+1k = \alpha + 1

Then

xα=xk1x^{\alpha} = x^{k - 1}

And the integral will be:

01xk1(1x)β1 dx\displaystyle\int_{0}^{1} x^{k - 1}(1 - x)^{\beta - 1}\ dx

This is just a beta function.

01xk1(1x)β1 dx=B(k,β)=Γ(k)Γ(β)Γ(k+β)\displaystyle\int_{0}^{1} x^{k - 1}(1 - x)^{\beta - 1}\ dx = B(k,\beta) = \frac{\Gamma(k)\Gamma(\beta)}{\Gamma(k + \beta)}

Or

B(k,β)=B(α+1,β)=Γ(α+1)Γ(β)Γ(α+β+1)\displaystyle B(k,\beta) = B(\alpha + 1,\beta) = \frac{\Gamma(\alpha + 1)\Gamma(\beta)}{\Gamma(\alpha + \beta + 1)}

To simplify the gamma function, visit this thread.


Go to post #5 written by professor fresh_42.
 
Nice try, but that was not the trick. The trick is:

xα=xα+11x^{\alpha} = x^{\alpha + 1 - 1}

Let k=α+1k = \alpha + 1

Then

xα=xk1x^{\alpha} = x^{k - 1}

And the integral will be:

01xk1(1x)β1 dx\displaystyle\int_{0}^{1} x^{k - 1}(1 - x)^{\beta - 1}\ dx

This is just a beta function.

01xk1(1x)β1 dx=B(k,β)=Γ(k)Γ(β)Γ(k+β)\displaystyle\int_{0}^{1} x^{k - 1}(1 - x)^{\beta - 1}\ dx = B(k,\beta) = \frac{\Gamma(k)\Gamma(\beta)}{\Gamma(k + \beta)}

Or

B(k,β)=B(α+1,β)=Γ(α+1)Γ(β)Γ(α+β+1)\displaystyle B(k,\beta) = B(\alpha + 1,\beta) = \frac{\Gamma(\alpha + 1)\Gamma(\beta)}{\Gamma(\alpha + \beta + 1)}

To simplify the gamma function, visit this thread.


Go to post #5 written by professor fresh_42.
thank mario99 very much

you explain it nicely

fresh_42 say Γ(x+1)=xΓ(x)\displaystyle \Gamma(x + 1) = x\Gamma(x)

so Γ(α+1)Γ(β)Γ(α+β+1)=αΓ(α)Γ(β)(α+β)Γ(α+β)\displaystyle \frac{\Gamma(\alpha + 1)\Gamma(\beta)}{\Gamma(\alpha + \beta + 1)} = \frac{\alpha\Gamma(\alpha)\Gamma(\beta)}{(\alpha + \beta)\Gamma(\alpha + \beta)}

still αΓ(α)Γ(β)(α+β)Γ(α+β)αα+β\displaystyle \frac{\alpha\Gamma(\alpha)\Gamma(\beta)}{(\alpha + \beta)\Gamma(\alpha + \beta)} \neq \frac{\alpha}{\alpha + \beta}☹️
 
Top