[
https://issues.apache.org/jira/browse/NUMBERS-167?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17706930#comment-17706930
]
Gilles Sadowski commented on NUMBERS-167:
-----------------------------------------
In the use case which I'm mentioning, the pre-computed value is use in a loop
that iterates at most 5 times (hard-coded). So the "LogGamma" value would be
needlessly evaluated at most 4 times (or once, if the code path doesn't use it).
Given that you
{quote}did not see a simple implementation
{quote}
and
{quote}decided it was non-trivial.
{quote}
my feeling is that the complexity of the API envisioned in the previous
discussion doesn't seem worth it...
Also, the "for" loop is within this function:
{code}
/** [...]
* This function returns an approximation of the inverse cumulative
* distribution function for the chi-square distribution. That is, given p
* and nu it computes the quantile x such that Pr{X2 ≤ x} = p, where X2
is
* a random variable from the chi-square distribution with nu degrees of
* freedom.
*
* NaN is returned if p < 0 or p > 1. Zero and positive infinity is
returned
* when p is equal to 0 and 1, respectively, or when p is too close to these
* values for a more accurate evaluation.
*
* This is a straight-forward implementation of the Fortran algorithm AS91
* (D.J. Best & D.E. Roberts, The percentage points of the Chi2
* distribution, Applied Statistics 24, 385, 1975), with some modifications
* to handle extreme values of p better.
*
* The accuracy is essentially determined by the accuracy of the lnGamma()
* and incompleteGammaP() functions used to compute the cumulative
* distribution.
*
* @param p
* probability (in the range 0 to 1)
* @param nu
* number of degrees of freedom
* @return lower tail quantile
*/
public static double chiSquareCdfInv(double p, double nu) {
// ...
}
{code}
> RegularizedGamma.P with precomputed LogGamma value
> --------------------------------------------------
>
> Key: NUMBERS-167
> URL: https://issues.apache.org/jira/browse/NUMBERS-167
> Project: Commons Numbers
> Issue Type: Wish
> Components: gamma
> Reporter: Gilles Sadowski
> Priority: Minor
> Fix For: 1.2
>
> Attachments: pr_106.patch
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> We have
> {code:java}
> double v = RegularizedGamma.P.value(a, x);
> {code}
> where method {{value}} internally calls {{LogGamma.value(a)}}.
> There is a use-case for
> {code:java}
> double logGammaA = LogGamma.value(a);
> double v = RegularizedGamma.P.value(a, x, logGammaA);
> {code}
> for when the user varies {{x}} but not {{a}}.
> Method name TBD: Another overload of {{value}} may be confusing (?).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
