The Commons Math team is pleased to announce the commons-math3-3.2-SNAPSHOT release!

The Math project is a library of lightweight, self-contained mathematics and statistics components addressing the most common practical problems not immediately available in the Java programming language or commons-lang.

Changes in this version include:

New features:
o All contents of package "o.a.c.m.stat.clustering" refactored into
        new package "o.a.c.m.ml.clustering" and added support for additional
        distance measures in package "o.a.c.m.ml.distance": "CanberraDistance",
        "ChebyshevDistance", "EuclideanDistance" and "ManhattanDistance".  Issue: MATH-917,MATH-918,MATH-919,MATH-920. Thanks to Reid Hochstedler. 
o Added Multivariate Normal Mixture Model Fitting by Expectation Maximization.  Issue: MATH-817. Thanks to Jared Becksfort. 
o Added a way to trigger only increasing or decreasing events in ODE integration.  Issue: MATH-811. 
o Added array-scaling methods to MathArrays.  Issue: MATH-946. Thanks to Jared Becksfort. 
o Added the Lévy distribution.  Issue: MATH-460. Thanks to Andrew Waterman. 
o Implementations for inverse error function and inverse complementary
        error functions have been added.  Issue: MATH-948. 
o Added utilities for prime numbers.  Issue: MATH-845. Thanks to Sébastien Riou. 
o Check bounds in multi-start vector optimizers.  Issue: MATH-914. 
o Added discrete distributions.  Issue: MATH-941. Thanks to Piotr Wydrych. 
o Added Hermite interpolator for RealFieldElement instances. 
o Added RealFieldElement interface to represent anything that is
        real number like, implemented by both Decimal64, Dfp and DerivativeStructure. 
o Added partial derivatives computation for 3D vectors and rotations. 
o Added accurate linear combination of DerivativeStructure instances,
        avoiding cancellation. 
o Added conversion of gradients and Hessians from spherical to Cartesian
        coordinates in 3D. 

Fixed Bugs:
o Fixed inverse cumulative probability for uniform distribution.  Issue: MATH-957. Thanks to Evan Ward. 
o "SpearmansCorrelation" now works correctly in case of a provided
        "NaturalRanking" with a "NaNStrategy.REMOVED" strategy and the input
        data contains NaN values. From version 4.0 onwards this strategy will
        not be supported anymore.  Issue: MATH-891. 
o Fixed Complex.reciprocal() for zero argument.  Issue: MATH-934. 
o AbstractRealMatrix will now check for rectangular input arrays in
        its copySubMatrix methods.  Issue: MATH-862. 
o Increment iteration counter in optimization algorithms.  Issue: MATH-949. 
o Fixed missing update in ODE event handlers, when a RESET_STATE is triggered.  Issue: MATH-950. 
o Fixed infinite loop when NaN occurs in singular value decomposition.  Issue: MATH-947. 
o Extended ranges for FastMath performance tests.  Issue: MATH-580. 
o Finalized implementation of diagonal matrix.  Issue: MATH-925. 
o Added rank revealing QR decomposition.  Issue: MATH-630. Thanks to Christopher Nix. 
o ArrayFieldVector can now be constructed from any FieldVector.  Issue: MATH-570. Thanks to Arne Plöse. 
o Improved checking of null vector elements.  Issue: MATH-861. Thanks to Sébastien Brisard. 
o Fixed generation of long random numbers between two bounds.  Issue: MATH-936. 
o Fixed creation of generic array.  Issue: MATH-942. Thanks to Piotr Wydrych. 
o Fixed abstract test class naming that broke ant builds.  Issue: MATH-940. Thanks to Piotr Wydrych. 
o Allow covariance to be computed for one-dimensional variables.  Issue: MATH-939. Thanks to Piotr Wydrych. 
o Fixed accuracy of 3D Line.revert().  Issue: MATH-938. 
o Improved javadoc to explain how switching functions should
        behave across events in ODE events detection.  Issue: MATH-937. 
o Fixed DerivativeStructure.atan2 for special cases when both arguments are +/-0.  Issue: MATH-935. 
o Improved class javadoc wrt convergence criteria and added 
        additional constructors to override the default epsilon and cut-off
        values in class "SimplexSolver".  Issue: MATH-930. 
o Fixed truncated value in "MultivariateNormalDistribution".  Issue: MATH-929. Thanks to Piotr Wydrych. 
o Made "BitStreamGenerator" implement the "Serializable" interface.  Issue: MATH-927. Thanks to Dennis Hendriks. 

Changes:
o Replaced hard-coded numbers in "LevenbergMarquardtOptimizer".  Issue: MATH-956. 
o Fixed loading of test file when path contains a space.  Issue: MATH-955. Thanks to Evan Ward. 
o Improved speed of FastMath.abs methods for all signatures, by removing branching.  Issue: MATH-954. Thanks to Charles Cooper. 
o Improved speed of several FastMath methods.  Issue: MATH-953. Thanks to Charles Cooper. 
o Improved speed of FastMath copysign methods.  Issue: MATH-951. Thanks to Charles Cooper. 
o Made EmpiricalDisribution smoothing kernel pluggable.  Issue: MATH-671. 
o Allow direct use of SummaryStatistics in one-way ANOVA.  Issue: MATH-877. Thanks to Peter Andrews. 
o Normal distribution now uses a direct implementation of the
        inverse error function to compute inverse cumulative probability
        instead of relying on a numerical solver. This is much faster,
        more accurate and does not need convergence threshold. 
o Throw "MathUnsupportedOperationException" from optimizers that do
        not support constraints (previous behaviour was to silently ignore
        the "SimpleBounds" argument).  Issue: MATH-933. 
o Greater efficiency in "UnitSphereRandomVectorGenerator".  Issue: MATH-931. Thanks to Sean Owen. 


Have fun!
-Commons Math team