2
$\begingroup$

I obtained the following integral when looking for a probability density function: $$\int_0^1 x^{\alpha-1} \,(1-x) ^{-A}\, {}_2F_1 (1-A, \alpha -1-A, \alpha -A, x) \,dx$$ Can anyone please give me some hints of evaluating the value of the integral?

$\endgroup$
3
  • $\begingroup$ You should tell us what is known about $\alpha$ and $A$. Are they positive? In a given range? $\endgroup$ Commented Aug 21, 2013 at 10:59
  • $\begingroup$ Yes. $A$ and $\alpha$ are both positive. $\endgroup$
    – user38925
    Commented Aug 22, 2013 at 2:47
  • $\begingroup$ you will also need A<1 to make the integral converge $\endgroup$ Commented Aug 22, 2013 at 7:08

1 Answer 1

1
$\begingroup$

If $\Re(A)<1$ and $\Re(\alpha)>0$ Mathematica gives:

$$\int_0^1 \mbox{ }x^{\alpha-1} \,(1-x) ^{-A}\, {}_2F_1 (1-A, \alpha -1-A; \alpha -A; x) \mbox{ }dx$$ $$\mbox{ }=\frac{\Gamma(\alpha)\Gamma(1-A)}{\Gamma(1+\alpha-A)}\mbox{ } \mbox{}_3F_2(\alpha,1-A,\alpha-A-1;\alpha-A,1+\alpha-A;1)$$

$\endgroup$
3
  • 1
    $\begingroup$ This is what you get by integrating termwise, using the beta integral. $\endgroup$
    – Ira Gessel
    Commented Aug 21, 2013 at 12:44
  • $\begingroup$ @IraGessel By termwise, I assume you mean "integrating the hypergeometric series termwise"?! $\endgroup$
    – Igor Rivin
    Commented Aug 21, 2013 at 17:34
  • $\begingroup$ Yes, I mean integrating $x^{\alpha -1}(1-x)^{-A}$ times each term of the hypergeometric series. $\endgroup$
    – Ira Gessel
    Commented Aug 21, 2013 at 19:10

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .