metanumerics Releases Rss Feedhttp://metanumerics.codeplex.com/Release/ProjectReleases.aspxmetanumerics Releases Rss DescriptionUpdated Release: Version 3.1.0 (Dec 26, 2016)https://metanumerics.codeplex.com/releases/view/630182<div class="wikidoc"><ul><li>Functions: We have added more advanced functions, including the Jacobi elliptic functions and Stirling numbers.</li></ul>
<ul><li>Analysis: We have added methods for integrating ordinary differential equations. We handle both simple and multiple, coupled ODEs, and provide special methods for conservative ODEs. Our ODE integration framework allows you to specify a Listener method which receives updates as the ODE is integrated.</li></ul>
<ul><li>Statistics: We have added time series analysis using the new TimeSeries class. In addition to computing power spectra and sample and population autocovariance estimates, we can difference and fit time series to AR(1) and MA(1) models. We have also added (balanced) two-way ANOVAs and multi-linear logistic regression. Our linear regression now returns a rich result object which makes it easy to access R and F tests, an ANOVA sum-of-squares analysis, get residuals, and obtain predictions with uncertainties for new points. We now support more than 25 common and not-so-common probability distributions.</li></ul>
<ul><li>We have made several performance enhancements and bug fixes.</li></ul>
<br /><a href="https://www.nuget.org/packages/Meta.Numerics/">Nuget</a> is the easiest way to use the package nowadays, but MSI and ZIP installation are still supported.</div><div class="ClearBoth"></div>ichbinTue, 27 Dec 2016 01:57:55 GMTUpdated Release: Version 3.1.0 (Dec 26, 2016) 20161227015755AReleased: Version 3.1.0 (Dec 26, 2016)http://metanumerics.codeplex.com/releases/view/630182
<div class="wikidoc">
<ul>
<li>Functions: We have added more advanced functions, including the Jacobi elliptic functions and Stirling numbers.</li></ul>
<ul>
<li>Analysis: We have added methods for integrating ordinary differential equations. We handle both simple and multiple, coupled ODEs, and provide special methods for conservative ODEs. Our ODE integration framework allows you to specify a Listener method which
receives updates as the ODE is integrated.</li></ul>
<ul>
<li>Statistics: We have added time series analysis using the new TimeSeries class. In addition to computing power spectra and sample and population autocovariance estimates, we can difference and fit time series to AR(1) and MA(1) models. We have also added
(balanced) two-way ANOVAs and multi-linear logistic regression. Our linear regression now returns a rich result object which makes it easy to access R and F tests, an ANOVA sum-of-squares analysis, get residuals, and obtain predictions with uncertainties for
new points. We now support more than 25 common and not-so-common probability distributions.</li></ul>
<ul>
<li>We have made several performance enhancements and bug fixes.</li></ul>
<br>
<a href="https://www.nuget.org/packages/Meta.Numerics/">Nuget</a> is the easiest way to use the package nowadays, but MSI and ZIP installation are still supported.</div>
<div></div>
Tue, 27 Dec 2016 01:57:55 GMTReleased: Version 3.1.0 (Dec 26, 2016) 20161227015755AUpdated Release: Version 3.0.0 (Feb 11, 2015)https://metanumerics.codeplex.com/releases/view/611568<div class="wikidoc"><ul><li>Functions: More integer, real, and complex advanced functions have been added. These include the complex Riemann zeta function, Clausen and polylog functions, and Bell, Bernoulli, and Fibonacci numbers. We have introduced a Permutation object along with methods that allow permutations to be generated, analyzed, multiplied, and applied. The spin objects and associated classes have been moved into this namespace.</li></ul>
<ul><li>Analysis: Classes and methods for the numerical analysis of user-supplied functions have been moved into this namespace. Our algorithms for multi-dimensional integration and optimization have been updated; they are becoming competitive with state-of-the-art dedicated libraries. We have added methods for global as well as local multi-dimensional optimization. Most analysis methods now return EvaluationResult objects that contain more information than a simple value.</li></ul>
<ul><li>Statistics: We have introduced a histogram class for analyzing discrete or binned data. We have updated several tests which used to use asymptotic null distributions to compute P-values to instead use exact or more accurate approximations to small-sample distributions. We have changed our maximum likelihood fit method to accept a simple factory delegate; it is no longer required that the distribution to be fit implement the IParameterizedDistribution interface.</li></ul>
<ul><li>Distributions: We have broken out properties common to both discrete and continuous univariate distributions into a parent UnivariateDistribution class. We have added cumulant methods for all distributions. We have introduced a MomentMath class that can convert between raw moments, central moments, and cumulants.</li></ul>
<ul><li>Matrices: We have implemented read-only matrices and vectors. These are now returned by many APIs whose result is a matrix or vector. Since we do not have to worry that the caller might change the returned object, we can simply return a reference to the underlying stored data without have to take the time and memory to copy it. Of course, if you do need to change the object you can simply call Copy and not have paid any higher price in time or memory than you did previously.</li></ul>
<ul><li>Core: We have introduced additional convenience functions in the MoreMath class, including trigonometric functions that are computed to full precision for all arguments.</li></ul>
<ul><li>Meta.Numerics is now a .NET portable class library (PCL), meaning the same DLL can be used by the desktop applications (framework version 4.0 and above), windows store apps, windows phone apps (version 7.5 and above), and Silverlight apps.</li></ul>
<ul><li>Since this is a major version bump, we have allowed ourselves to make breaking changes in ways that we believe improve the API organization. All previous functionality still exists. Mostly we have moved classes between namespaces, or methods between classes, or changed methods to properties. You should be able to quickly adapt most code built against 2.2.0 to use 3.0.0.</li></ul></div><div class="ClearBoth"></div>ichbinThu, 12 Feb 2015 06:38:01 GMTUpdated Release: Version 3.0.0 (Feb 11, 2015) 20150212063801AReleased: Version 3.0.0 (Feb 11, 2015)http://metanumerics.codeplex.com/releases/view/611568
<div class="wikidoc">
<ul>
<li>Functions: More integer, real, and complex advanced functions have been added. These include the complex Riemann zeta function, Clausen and polylog functions, and Bell, Bernoulli, and Fibonacci numbers. We have introduced a Permutation object along with
methods that allow permutations to be generated, analyzed, multiplied, and applied. The spin objects and associated classes have been moved into this namespace.</li></ul>
<ul>
<li>Analysis: Classes and methods for the numerical analysis of user-supplied functions have been moved into this namespace. Our algorithms for multi-dimensional integration and optimization have been updated; they are becoming competitive with state-of-the-art
dedicated libraries. We have added methods for global as well as local multi-dimensional optimization. Most analysis methods now return EvaluationResult objects that contain more information than a simple value.</li></ul>
<ul>
<li>Statistics: We have introduced a histogram class for analyzing discrete or binned data. We have updated several tests which used to use asymptotic null distributions to compute P-values to instead use exact or more accurate approximations to small-sample
distributions. We have changed our maximum likelihood fit method to accept a simple factory delegate; it is no longer required that the distribution to be fit implement the IParameterizedDistribution interface.</li></ul>
<ul>
<li>Distributions: We have broken out properties common to both discrete and continuous univariate distributions into a parent UnivariateDistribution class. We have added cumulant methods for all distributions. We have introduced a MomentMath class that can
convert between raw moments, central moments, and cumulants.</li></ul>
<ul>
<li>Matrices: We have implemented read-only matrices and vectors. These are now returned by many APIs whose result is a matrix or vector. Since we do not have to worry that the caller might change the returned object, we can simply return a reference to the
underlying stored data without have to take the time and memory to copy it. Of course, if you do need to change the object you can simply call Copy and not have paid any higher price in time or memory than you did previously.</li></ul>
<ul>
<li>Core: We have introduced additional convenience functions in the MoreMath class, including trigonometric functions that are computed to full precision for all arguments.</li></ul>
<ul>
<li>Meta.Numerics is now a .NET portable class library (PCL), meaning the same DLL can be used by the desktop applications (framework version 4.0 and above), windows store apps, windows phone apps (version 7.5 and above), and Silverlight apps.</li></ul>
<ul>
<li>Since this is a major version bump, we have allowed ourselves to make breaking changes in ways that we believe improve the API organization. All previous functionality still exists. Mostly we have moved classes between namespaces, or methods between classes,
or changed methods to properties. You should be able to quickly adapt most code built against 2.2.0 to use 3.0.0.</li></ul>
</div>
<div></div>
Thu, 12 Feb 2015 06:38:00 GMTReleased: Version 3.0.0 (Feb 11, 2015) 20150212063800AUpdated Release: Version 2.2.0 (Aug 12, 2013)https://metanumerics.codeplex.com/releases/view/110664<div class="wikidoc">What's new in 2.2.0:<br />
<ul><li>Multidimensional integration.</li>
<li>Polynomial interpolation, arithmetic, and calculus.</li>
<li>Faster non-uniform deviates and probit functions.</li>
<li>Exact Kendall, Kolmogorov-Smirnov, Kuiper, and Spearmann null distributions for small sample sizes.</li>
<li>Fit samples to Gamma and Lognormal distributions. Improved performance of other FitToSample methods.</li>
<li>Matrix powers and utility methods.</li></ul></div><div class="ClearBoth"></div>ichbinMon, 12 Aug 2013 10:50:03 GMTUpdated Release: Version 2.2.0 (Aug 12, 2013) 20130812105003AReleased: Version 2.2.0 (Aug 12, 2013)http://metanumerics.codeplex.com/releases/view/110664
<div class="wikidoc">What's new in 2.2.0:<br>
<ul>
<li>Multidimensional integration. </li><li>Polynomial interpolation, arithmetic, and calculus. </li><li>Faster non-uniform deviates and probit functions. </li><li>Exact Kendall, Kolmogorov-Smirnov, Kuiper, and Spearmann null distributions for small sample sizes.
</li><li>Fit samples to Gamma and Lognormal distributions. Improved performance of other FitToSample methods.
</li><li>Matrix powers and utility methods.</li></ul>
</div>
<div></div>
Mon, 12 Aug 2013 10:50:02 GMTReleased: Version 2.2.0 (Aug 12, 2013) 20130812105002AUpdated Release: Version 2.1.0 (Sep 04, 2012)http://metanumerics.codeplex.com/releases/view/93895<div class="wikidoc">New features in this release include:<br />
<ul><li>Improved FFT. Fourier transforms are fast for all lengths, even those with large prime factors.</li>
<li>More advanced functions: incomplete elliptic integrals of the second kind, the polygamma function, and complex exponential integrals. New Bessel function methods simultaneously return (J, J', Y, Y') or (I, I', K, K'). New advanced integer functions include primality testing and harmonic numbers.</li>
<li>SpareSquareMatrix now supports the iterative solution of large systems of equations.</li>
<li>New statistical analysis features include sample transformations, polynomial regression for bivariate samples, and fitting samples to beta distributions.</li></ul>
<br />This release also fixes a bug in our SVD algorithm which occured for some matrices with zero singular values, a bug in the Sample class's inverse CDF (i.e. percentile to score) method, and several other bugs.</div><div class="ClearBoth"></div>ichbinTue, 04 Sep 2012 18:26:44 GMTUpdated Release: Version 2.1.0 (Sep 04, 2012) 20120904062644PReleased: Version 2.1.0 (Sep 04, 2012)http://metanumerics.codeplex.com/releases/view/93895
<div class="wikidoc">New features in this release include:<br>
<ul>
<li>Improved FFT. Fourier transforms are fast for all lengths, even those with large prime factors.
</li><li>More advanced functions: incomplete elliptic integrals of the second kind, the polygamma function, and complex exponential integrals. New Bessel function methods simultaneously return (J, J', Y, Y') or (I, I', K, K'). New advanced integer
functions include primality testing and harmonic numbers. </li><li>SpareSquareMatrix now supports the iterative solution of large systems of equations.
</li><li>New statistical analysis features include sample transformations, polynomial regression for bivariate samples, and fitting samples to beta distributions.</li></ul>
<br>
This release also fixes a bug in our SVD algorithm which occured for some matrices with zero singular values, a bug in the Sample class's inverse CDF (i.e. percentile to score) method, and several other bugs.</div>
<div></div>
Tue, 04 Sep 2012 18:26:43 GMTReleased: Version 2.1.0 (Sep 04, 2012) 20120904062643PUpdated Release: Version 2.0.0 (Apr 06, 2011)http://metanumerics.codeplex.com/releases/view/63906<div class="wikidoc">There is a <i>ton</i> of new and improved functionality in this major release:
<ul><li>Singular Value Decomposition (SVD), sparse matrices, matrix norms, and a new matrix hierarchy</li>
<li>Fast Fourier Transforms (FFT), including non-power-of-2 lengths</li>
<li>Principal Component Analysis (PCA), logistic regression, Kruskal-Wallis test</li>
<li>Easier to load data from databases or Excel spreadsheets</li>
<li>More distributions: Gamma, Pearson, discrete uniform, discrete geometric</li>
<li>Faster, dedicated methods to fit distributions to samples and generate nonuniform random deviates</li></ul>
<br />Because this is a major release, we took the oportunity to improve our API in ways that may not be backwards-compatible (see our <a href="http://metanumerics.codeplex.com/wikipage?title=Versioning">Versioning policy</a>). We did not, however, remove any functionality. Any previous code does not compile against this new version should be easy to adapt; if you have any difficulties you can't immediately figure out, just ask in our <a href="http://metanumerics.codeplex.com/wikipage?title=http%3a%2f%2fmetanumerics.codeplex.com%2fdiscussions">Discussion Forum</a>.</div><div class="ClearBoth"></div>ichbinWed, 06 Apr 2011 08:45:18 GMTUpdated Release: Version 2.0.0 (Apr 06, 2011) 20110406084518AReleased: Version 2.0.0 (Apr 06, 2011)http://metanumerics.codeplex.com/releases/view/63906
<div class="wikidoc">There is a <i>ton</i> of new and improved functionality in this major release:
<ul>
<li>Singular Value Decomposition (SVD), sparse matrices, matrix norms, and a new matrix hierarchy
</li><li>Fast Fourier Transforms (FFT), including non-power-of-2 lengths </li><li>Principal Component Analysis (PCA), logistic regression, Kruskal-Wallis test </li><li>Easier to load data from databases or Excel spreadsheets </li><li>More distributions: Gamma, Pearson, discrete uniform, discrete geometric </li><li>Faster, dedicated methods to fit distributions to samples and generate nonuniform random deviates</li></ul>
<br>
Because this is a major release, we took the oportunity to improve our API in ways that may not be backwards-compatible (see our
<a href="http://metanumerics.codeplex.com/wikipage?title=Versioning">Versioning policy</a>). We did not, however, remove any functionality. Any previous code does not compile against this new version should be easy to adapt; if you have any difficulties you
can't immediately figure out, just ask in our <a href="http://metanumerics.codeplex.com/wikipage?title=http%3a%2f%2fmetanumerics.codeplex.com%2fdiscussions">
Discussion Forum</a>.</div>
<div></div>
Wed, 06 Apr 2011 08:45:18 GMTReleased: Version 2.0.0 (Apr 06, 2011) 20110406084518AUpdated Release: Version 1.5.0 (Sep 30, 2010)http://metanumerics.codeplex.com/releases/view/52952<div class="wikidoc">Meta.Numerics is a .NET library for scientific computation.<br /><br />Notable new features in this release include:<br />
<ul><li>More speical functions: associated Legendre and Laguerre polynomials, elliptic integrals, integer partitions</li>
<li>Faster matrix multiplication, inversion, LU and QR decomposition</li>
<li>More statistical tests: Z test, sign test, one-way ANOVA</li>
<li>More probability distributions, including discrete distributions like Binomial and Poisson</li>
<li>Bug-fixes and improved documentation</li></ul></div><div class="ClearBoth"></div>ichbinSat, 02 Oct 2010 06:24:21 GMTUpdated Release: Version 1.5.0 (Sep 30, 2010) 20101002062421AReleased: Version 1.5.0 (Sep 30, 2010)http://metanumerics.codeplex.com/releases/view/52952<div class=wikidoc>Meta.Numerics is a .NET library for scientific computation.<br><br>Notable new features in this release include:<br>
<ul><li>More speical functions: associated Legendre and Laguerre polynomials, elliptic integrals, integer partitions</li>
<li>Faster matrix multiplication, inversion, LU and QR decomposition</li>
<li>More statistical tests: Z test, sign test, one-way ANOVA</li>
<li>More probability distributions, including discrete distributions like Binomial and Poisson</li>
<li>Bug-fixes and improved documentation</li></ul></div><div></div>Sat, 02 Oct 2010 06:24:21 GMTReleased: Version 1.5.0 (Sep 30, 2010) 20101002062421AUpdated Release: Version 1.5.0 (Sep 30, 2010)http://metanumerics.codeplex.com/releases/view/52952<div class="wikidoc">Meta.Numerics is a .NET library for scientific computation.<br /><br />Notable new features in this release include:<br />
<ul><li>More speical functions: associated Legendre and Laguerre polynomials, elliptic integrals, integer patitions</li>
<li>Faster matrix multiplication, inversion, LU and QR decomposition</li>
<li>More statistical tests: Z test, sign test, one-way ANOVA</li>
<li>More probability distributions, including discrete distributions like Binomial and Poisson</li>
<li>Bug-fixes and improved documentation</li></ul></div><div class="ClearBoth"></div>ichbinThu, 30 Sep 2010 17:18:51 GMTUpdated Release: Version 1.5.0 (Sep 30, 2010) 20100930051851PReleased: Version 1.5.0 (Sep 30, 2010)http://metanumerics.codeplex.com/releases/view/52952<div class=wikidoc>Meta.Numerics is a .NET library for scientific computation.<br><br>Notable new features in this release include:<br>
<ul><li>More speical functions: associated Legendre and Laguerre polynomials, elliptic integrals, integer patitions</li>
<li>Faster matrix multiplication, inversion, LU and QR decomposition</li>
<li>More statistical tests: Z test, sign test, one-way ANOVA</li>
<li>More probability distributions, including discrete distributions like Binomial and Poisson</li>
<li>Bug-fixes and improved documentation</li></ul></div><div></div>Thu, 30 Sep 2010 17:18:51 GMTReleased: Version 1.5.0 (Sep 30, 2010) 20100930051851PUpdated Release: Version 1.5.0 (Sep 30, 2010)http://metanumerics.codeplex.com/releases/view/52952<div class="wikidoc">Meta.Numerics is a .NET library for scientific computation.<br /><br />Notable new features in this release include:<br />
<ul><li>More speical functions: associated Legendre and Laguerre polynomials, elliptic integrals, integer patitions</li>
<li>Faster matrix multiplication, inversion, LU and QR decomposition</li>
<li>More statistical tests: Z test, sign test, one-way ANOVA</li>
<li>More probability distributions, including discrete distributions like Binomial and Poisson</li>
<li>Bug-fixes and improved documentation</li></ul></div><div class="ClearBoth"></div>ichbinThu, 30 Sep 2010 17:18:26 GMTUpdated Release: Version 1.5.0 (Sep 30, 2010) 20100930051826PUpdated Release: Version 1.4.0 (May 05, 2010)http://metanumerics.codeplex.com/releases/view/44777<div class="wikidoc">Meta.Numerics is a .NET library for scientific computation.<br /><br />New features in this release:
<ul><li>Many new advanced functions: modified Bessel functions, Airy function, Coulomb wave functions, Dilogarithm function.</li>
<li>Linear regression of multivariate data sets, with covariances and a F-tested significance.</li>
<li>6j symbols and other new spinor math functions.</li>
<li>Triangular and beta distributions.</li>
<li>Improved documentation.</li>
<li>Bugfixes to chi squared inverse PDF function and complex tanget function.</li></ul></div><div class="ClearBoth"></div>ichbinTue, 11 May 2010 07:23:25 GMTUpdated Release: Version 1.4.0 (May 05, 2010) 20100511072325AReleased: Version 1.4.0 (May 05, 2010)http://metanumerics.codeplex.com/releases/view/44777<div>Meta.Numerics is a .NET library for scientific computation.<br><br>New features in this release:
<ul><li>Many new advanced functions: modified Bessel functions, Airy function, Coulomb wave functions, Dilogarithm function.</li>
<li>Linear regression of multivariate data sets, with covariances and a F-tested significance.</li>
<li>6j symbols and other new spinor math functions.</li>
<li>Triangular and beta distributions.</li>
<li>Improved documentation.</li>
<li>Bugfixes to chi squared inverse PDF function and complex tanget function.</li></ul></div><div></div>Tue, 11 May 2010 07:23:25 GMTReleased: Version 1.4.0 (May 05, 2010) 20100511072325AUpdated Release: Version 1.4.0 (May 05, 2010)http://metanumerics.codeplex.com/releases/view/44777<div class="wikidoc">Meta.Numerics is a .NET library for scientific computation.<br /><br />New features in this release:
<ul><li>Many new advanced functions: modified Bessel functions, Airy function, Coulomb wave functions, Dilogarithm function.</li>
<li>Linear regression of multivariate data sets, with covariances and a F-tested significance.</li>
<li>6j symbols and other new spinor math functions.</li>
<li>Triangular and beta distributions.</li>
<li>Improved documentation.</li>
<li>Bugfixes to chi squared inverse PDF function and complex tanget function.</li></ul></div><div class="ClearBoth"></div>ichbinWed, 05 May 2010 17:41:46 GMTUpdated Release: Version 1.4.0 (May 05, 2010) 20100505054146PReleased: Version 1.4.0 (May 05, 2010)http://metanumerics.codeplex.com/releases/view/44777<div>Meta.Numerics is a .NET library for scientific computation.<br><br>New features in this release:
<ul><li>Many new advanced functions: modified Bessel functions, Airy function, Coulomb wave functions, Dilogarithm function.</li>
<li>Linear regression of multivariate data sets, with covariances and a F-tested significance.</li>
<li>6j symbols and other new spinor math functions.</li>
<li>Triangular and beta distributions.</li>
<li>Improved documentation.</li>
<li>Bugfixes to chi squared inverse PDF function and complex tanget function.</li></ul></div><div></div>Wed, 05 May 2010 17:41:46 GMTReleased: Version 1.4.0 (May 05, 2010) 20100505054146PUpdated Release: Version 1.4.0 (May 05, 2010)http://metanumerics.codeplex.com/releases/view/44777<div class="wikidoc">Meta.Numerics is a .NET library for scientific computation.<br /><br />New features in this release:
<ul><li>Many new advanced functions: modified Bessel functions, Airy function, Coulomb wave functions, Dilogarithm function.</li>
<li>Linear regression of multivariate data sets, with covariances and a F-tested significance.</li>
<li>6j symbols and other new spinor math functions.</li>
<li>Triangular and beta distributions.</li>
<li>Improved documentation.</li>
<li>Bugfixes to chi squared inverse PDF function and complex tanget function.</li></ul></div><div class="ClearBoth"></div>ichbinWed, 05 May 2010 17:07:46 GMTUpdated Release: Version 1.4.0 (May 05, 2010) 20100505050746P