[
https://issues.apache.org/jira/browse/STATISTICS-96?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18108755#comment-18108755
]
Alex Herbert commented on STATISTICS-96:
----------------------------------------
NaN parameters will result in NaN results from the distribution probability
methods. These are always invalid. Checks for NaN parameters have been added in
commit:
9ee1a60ab16176c75876ba671b7bcd1de96431a9
> NaN (and often Infinity) parameters are accepted in many distribution
> factories; location parameters are unvalidated
> --------------------------------------------------------------------------------------------------------------------
>
> Key: STATISTICS-96
> URL: https://issues.apache.org/jira/browse/STATISTICS-96
> Project: Commons Statistics
> Issue Type: Bug
> Components: distribution
> Affects Versions: 1.3
> Reporter: Alex Herbert
> Priority: Minor
>
> Use of NaN and infinite parameters when constructing a distribution results
> in an invalid distribution. Parameter validation should be added to guard
> incorrect usage.
> The following distributions require updating:
> * BetaDistribution
> * CauchyDistribution
> * DistributionException
> * ExponentialDistribution
> * FDistribution
> * FoldedNormalDistribution
> * GammaDistribution
> * GeometricDistribution
> * GumbelDistribution
> * LaplaceDistribution
> * LevyDistribution
> * LogNormalDistribution
> * LogisticDistribution
> * NakagamiDistribution
> * NormalDistribution
> * ParetoDistribution
> * PascalDistribution
> * PoissonDistribution
> * TDistribution
> * TrapezoidalDistribution
> * TriangularDistribution
> * TruncatedNormalDistribution
> * WeibullDistribution
> * ZipfDistribution
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
