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.

We would like to find an upper bound on the following function:

$\left(\frac{\omega_1}{\omega_2}\right)^{(\alpha_1-1)} \frac{\Gamma(\alpha_1+\alpha_2-1)}{\Gamma(\alpha_1)\Gamma(\alpha_2)} {}_2F_1\left(\alpha_1-1,\alpha_1+\alpha_2-1,\alpha_1,-\frac{\omega_1}{\omega_2}\right)$

for all $\omega_1,\omega_2 > 0$ and $\alpha_1,\alpha_2 >1$.

From the way we derived this function, we know that it is smaller than 1, but we would like to find an upper bound (expressed in the parameters $\omega_1,\omega_2,\alpha_1,\alpha_2$) that is as tight as possible.

Thank you in advance for all possible suggestions and insights!

Sebas

share|improve this question
add comment

1 Answer 1

up vote 2 down vote accepted

This is the regularized incomplete beta function $I_x(a,b)=B(x;a,b)/B(a,b)$ for $$ x=\omega_1/(\omega_1+\omega_2),\qquad a=\alpha_1-1,\qquad b=\alpha_2. $$ Recall that $B(a,b)=B(1;a,b)$ with $$ B(x;a,b) = \int_0^x t^{a-1}\,(1-t)^{b-1}\,\mathrm{d}t. $$ Note that $0\le I_x(a,b)\le 1$ or every $0\le x\le 1$ and that the function $x\mapsto I_x(a,b)$ increases from $I_0(a,b)=0$ to $I_1(a,b)=1$. Furthermore, when $a$ et $b$ are positive integers, this is the cumulative distribution function of a random variable with binomial distribution, an interpretation which may be used to estimate $I_x(a,b)$. More precisely, $I_x(a,b)=P(X\le b-1)$ where the distribution of $X$ is binomial $(a+b-1,1-x)$.

See here. (Not knowing what the OP does and does not know, it is difficult to determine what to add to the present answer.)

For example, if $\alpha_1\to+\infty$ and $\alpha_2\to+\infty$ with $\alpha_1/\alpha_2\to\varrho$, and if $\omega_1$ and $\omega_2$ are fixed, then by the law of large numbers the limit of the quantity written by the OP is $1$ if $\omega_1>\varrho\omega_2$ and $0$ if $\omega_1<\varrho\omega_2$.

share|improve this answer
    
Thanks for the answer! It helped a lot :) –  user8991 Apr 5 '11 at 9:45
add comment

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.