Readme

Release Notes

Changes in IMSL C# Numerical Library Version 6.5.2

Updated to support .NET Framework version 4.5.2 and 4.5.1

MKL is not supported

Flexnet license management has been removed

`Imsl.DataMining.Neural`
	Neural Nets
		`LeastSquaresTrainer.Train`	Improved calculation of the gradient.

Changes in IMSL C# Numerical Library Version 6.5.1

`Imsl.Math`
	Optimization
		`MinConNLP.Solve`	Resolved buffer overrun issue.

Classes and Methods New to the IMSL C# Numerical Library Version 6.5

`Imsl.Math`
	Error Handling
		`NoProgressException`	Exception thrown when More's technique is not making any progress.
	Linear Systems
		`ComplexLU.GetL`	Returns the lower triangular portion of the LU factorization of input matrix "a".
		`ComplexLU.GetPermutationMatrix`	Returns the the permutation matrix which results from the LU factorization of input matrix "a".
		`ComplexLU.GetU`	Returns the unit upper triangular portion of the LU factorization of input matrix "a".
		`LU.GetL`	Returns the lower triangular portion of the LU factorization of input matrix "a".
		`LU.GetPermutationMatrix`	Returns the the permutation matrix which results from the LU factorization of input matrix "a".
		`LU.GetU`	Returns the unit upper triangular portion of the LU factorization of input matrix "a".
	Optimization
		`BoundedLeastSquares.Solve`	Solves a nonlinear least-squares problem subject to bounds on the variables using a modified Levenberg-Marquardt algorithm. This method is the identical replacement for the deprecated method BoundedLeastSquares.solve.
`Imsl.Stat`
	Error Handling
		`NoProgressException`	Exception thrown when More's technique is not making any progress.

Changes in IMSL C# Numerical Library Version 6.5

`General`
			Changed links to Microsoft SDK webpages for HTML documentation.
			Added support for the FLEXLM_BATCH FlexNet environment variable. This prevents popups from being displayed for FlexNet errors/warnings.
			The properties NumberOfProcessors and/or Parallel have been added to many of the classes/methods which use the Task Parallel Library in .NET 4.0.
`Gallery`
			Added TCB Spline to the Spline demo.
`Imsl.Chart2D`
	General
			Corrected problem with charts with axes on both the left and right side which resulted in inconsistencies of the tick intervals and ticks which appear to be outside the axis range.
`Imsl.Math`
	Linear Systems
		`SuperLU`	Made corrections so that the warning about a singular matrix is printed.
		`Cholesky`	Updated class description with A = RR^T.
		`Matrix.Multiply`	Overloaded method with processors argument.
		`ComplexMatrix.Multiply`	Overloaded method with processors argument.
	Optimization
		`BoundedLeastSquares`	Implemented a fix for a potential infinite loop in More's technique.
		`BoundedLeastSquares.solve`	This method has been deprecated and replaced by BoundedLeastSquares.Solve.
		`NonlinLeastSquares`	Updated the documentation for the methods and properties SetXscale, SetFscale, GradientTolerance, and StepTolerance.
			Implemented a fix for a potential infinite loop in More's technique.
		`QuadraticProgramming`	Modified the write statement in Example 3.
`Imsl.Stat`
	Regression
		`NonlinearRegression`	Implemented a fix for a potential infinite loop in More's technique.
	Probability Distribution Functions and Inverses
		`Pdf.NoncentralBeta`	Corrected an equation in the documentation.
		`GammaDistribution`	Enhanced the documentation.
		`IDistribution`	Enhanced the documentation.
		`IProbabilityDistribution`	Enhanced the documentation.
		`LogNormalDistribution`	Enhanced the documentation.
		`NormalDistribution`	Enhanced the documentation.
		`PoissonDistribution`	Enhanced the documentation.
	Random Number Generation
			Added description for the valid value of the seeds for the Random constructor.
		`Random.CanonicalCorrelation`	Changed from a static public method to a static method.
		`Random.NextGaussianCopula`	Overloaded version using the k argument has been deprecated.
		`Random.NextMultivariateNormal`	Overloaded version using the k argument has been deprecated.
		`Random.NextStudentsTCopula`	Overloaded version using the k argument has been deprecated.
	Probability Distribution Functions and Inverses
		`Cdf.Beta`	A new algorithm is used to improve accuracy in the tails of the distribution.
		`Cdf.NoncentralF`	Improved performance.
	Time Series and Forecasting
		`ARMA`	Removed extra terms from the difference equation in the documentation.
			Implemented a fix for a potential infinite loop in More's technique.

Classes and Methods New to the IMSL C# Numerical Library Version 6.0

`Imsl`
	Error Handling
		`Logger`	Logs intermediate results and notes from IMSL C# classes.
		`IMSLUnexpectedErrorException`	Signals that an unexpected error has occurred.
		`Messages.ThrowIllegalStateException`	Throws an `IllegalStateException` with a formatted String argument.
`Imsl.Math`
	Linear Systems
		`Matrix.InverseLowerTriangular`	Returns the inverse of the lower triangular matrix `a`.
		`Matrix.InverseUpperTriangular`	Returns the inverse of the upper triangular matrix `a`.
	Eigensystems Analysis
		`Eigen`	A constructor was added.
		`Eigen.MaxIterations`	Set or returns the maximum number of iterations.
		`Eigen.Solve`	Solves for the eigenvalues and (optionally) the eigenvectors of a real square matrix.
	Interpolation and Approximation
		`CsTCB`	Extension of the Spline class to handle a tension-continuity-bias (TCB) cubic spline, also known as a Kochanek-Bartels spline and is a generalization of the Catmull-Rom spline.
		`Spline2D.Integral`	Returns the value of an integral of a tensor-product spline on a rectangular domain.
		`Spline2DLeastSquares`	Computes a two-dimensional, tensor-product spline approximant using least squares.
	Differential Equations
		`ODE`	ODE represents and solves an initial-value problem for ordinary differential equations.
		`FeynmanKac`	Solves the generalized Feynman-Kac PDE.
		`OdeAdamsGear`	Extension of the ODE class to solve a stiff initial-value problem for ordinary differential equations using the Adams-Gear methods.
	Nonlinear Equations
		`ZerosFunction`	Finds the real zeros of a real, continuous, univariate function, f(x).
		`ZeroSystem.Logger`	Returns the logger object.
	Optimization
		`BoundedVariableLeastSquares`	Solve a linear least-squares problem with bounds on the variables.
		`NonNegativeLeastSquares`	Solves a linear least squares problem with nonnegativity constraints.
		`NumericalDerivatives`	Compute the Jacobian matrix for a function f(y) with m components in n independent variables.
	Special Functions
		`Sfun.Erfce`	Returns the exponentially scaled complementary error function.
		`Sfun.GammaIncomplete`	Evaluates the incomplete gamma function.
		`Sfun.Psi`	Returns the logarithmic derivative of the gamma function, also called the digamma function.
		`Sfun.Psi1`	Returns the ψ ₁ function, also known as the trigamma function.
`Imsl.Stat`
	Basic Statistics
	Regression
		`StepwiseRegression.Intercept`	Returns the intercept.
		`StepwiseRegression.SetMeans`	Sets the means of the variables.

	Analysis of Variance
		`ANCOVA`	Analyzes a one-way classification model with covariates.
		`ANOVA.GetConfidenceInterval`	Computes the confidence interval associated with the difference of means between two groups using a specified method.
	Time Series and Forecasting
		`AutoARIMA`	Automatically identifies time series outliers, determines parameters of a multiplicative seasonal model and produces forecasts that incorporate the effects of outliers whose effects persist beyond the end of the series.
		`ARMAOutlierIdentification`	Detects and determines outliers and simultaneously estimates the model parameters in a time series whose underlying outlier free series follows a general seasonal or nonseasonal ARMA model. Allows computation of forecasts.
		`LackOfFit`	Performs lack-of-fit test for a univariate time series or transfer function given the appropriate correlation function.
	Multivariate Analysis
		`DiscriminantAnalysis.Classify`	Classifies a set of observations using the linear or quadratic discriminant functions generated during the training process.
		`DiscriminantAnalysis.Downdate`	Removes a set of observations from the discriminant functions.
		`DiscriminantAnalysis.NumberOfRowsMissing`	Returns the number of rows of data encountered containing missing values (`Double.NaN`).
	Survival and Reliability Analysis
		`KaplanMeierEstimates`	Computes Kaplan-Meier (or product-limit) estimates of survival probabilities for a sample of failure times that possibly contain right consoring.
		`KaplanMeierECDF`	Computes the Kaplan-Meier reliability function estimates or the CDF based on failure data that may be multi-censored.
		`LifeTables`	Computes population (current) or cohort life tables based upon the observed population sizes at the middle (for population table) or the beginning (for cohort table) of some user specified age intervals.
		`ProportionalHazards`	Analyzes survival and reliability data using Cox's proportional hazards model.
	Probability Distribution Functions and Inverses
		`Cdf.NoncentralBeta`	Evaluates the noncentral beta cumulative distribution function (CDF).
		`Cdf.NoncentralF`	Evaluates the noncentral F cumulative distribution function (CDF).
		`Cdf.Logistic`	Evaluates the logistic cumulative probability distribution function.
		`Cdf.Pareto`	Evaluates the Pareto cumulative probability distribution function.
		`InvCdf.Logistic`	Returns the inverse of the logistic cumulative probability distribution function.
		`InvCdf.NoncentralBeta`	Evaluates the inverse of the noncentral beta cumulative distribution function (CDF).
		`InvCdf.NoncentralF`	Evaluates the inverse of the noncentral F cumulative distribution function (CDF).
		`InvCdf.Pareto`	Returns the inverse of the Pareto cumulative probability distribution function.
		`Pdf.Logistic`	Evaluates the logistic probability density function.
		`Pdf.NoncentralBeta`	Evaluates the noncentral beta probability density function (PDF).
		`Pdf.NoncentralChi`	Evaluates the noncentral chi-squared probability density function.
		`Pdf.NoncentralF`	Evaluates the noncentral F probability density function (PDF).
		`Pdf.NoncentralStudentsT`	Evaluates the noncentral Student's t probability density function.
		`Pdf.Normal`	Evaluates the normal (Gaussian) probability density function.
		`Pdf.Pareto`	Evaluates the Pareto probability density function.
		`IDistribution`	Public interface for the user-supplied distribution function.
		`IProbabilityDistribution`	Public interface for the user-supplied probability distribution function.
		`GammaDistribution`	Evaluates a gamma probability distribution.
		`LogNormalDistribution`	Evaluates a lognormal probability distribution.
		`NormalDistribution`	Evaluates a normal (Gaussian) probability distribution.
		`PoissonDistribution`	Evaluates a Poisson probability distribution.
	Random Number Generation
		`Random.CanonicalCorrelation`	Generates a canonical correlation matrix from an arbitrarily distributed multivariate deviate sequence with a Gaussian Copula dependence structure.
		`Random.NextGaussianCopula`	Generate pseudorandom numbers from a Gaussian Copula distribution.
		`Random.NextStudentsTCopula`	Generate pseudorandom numbers from a Student's t Copula distribution.
		`Random.NextZigguratNormalAR`	Generates pseudorandom numbers using the Ziggurat method.
`Imsl.Finance`
	Finance
		`BasisPart.GetDaysInYear`	Component of DayCountBasis. The day count basis consists of a month basis and a yearly basis. Each of these components implements this interface.
		`DayCountBasis.GetDaysInYear`	The Day Count Basis. Rules for computing the number or days between two dates or number of days in a year.
`Imsl.Chart`
	Chart 2D
		`Annotation`	Draws an annotation.
		`Draw.DrawClippedImage`	Draws an image such that any portion of the image beyond the axis range is clipped.
		`Treemap`	Treemap creates a chart from two arrays of double precision values or one data array and one array of java.awt.Color values.
`Imsl.Datamining`
	Data Mining
		`NaiveBayesClassifier`	Trains a Naive Bayes Classifier

Changes in IMSL C# Numerical Numerical Library Version 6.0

General

ErrorMessages Resource Bundle

Fixed misspelling, improved consistency and removed extra whitespace.

Gallery

The Gallery has been updated.

Optimization Chapter Introduction

Clarified documentation by stating that it is the responsibility of the user to ensure that the user-supplied evaluating function always returns valid results.

Quadrature Chapter Introduction

Clarified documentation by stating that it is the responsibility of the user to ensure that the user-supplied evaluating function always returns valid results.

Imsl.Math

Linear Systems

Matrix.Multiply

Performance increases can be realized by taking advantage of the MKL-enhanced version of the IMSL C# Numerical Library.

ComplexMatrix.Multiply

Performance increases can be realized by taking advantage of the MKL-enhanced version of the IMSL C# Numerical Library.

Eigensystem Analysis

Eigen

Performance increases can be realized by taking advantage of the MKL-enhanced version of the IMSL C# Numerical Library.

SymEigen

Performance increases can be realized by taking advantage of the MKL-enhanced version of the IMSL C# Numerical Library.

Interpolation and Approximation

CsShape

No longer issues an index out of bounds exception.

RadialBasis

Added examples for RadialBasis.Function, RadialBasis.Gaussian and RadialBasis.HardyMultiquadric to highlight the explicit use of different basis functions.

Documentation was added for RadialBasis.Gaussian and RadialBasis.HardyMultiquadric.

Transforms

FFT

Performance increases can be realized by taking advantage of the MKL-enhanced version of the IMSL C# Numerical Library.

ComplexFFT

Performance increases can be realized by taking advantage of the MKL-enhanced version of the IMSL C# Numerical Library.

Nonlinear Equations

ZeroFunction

This class has been deprecated. It has been replaced by the class ZerosFunction.

ZeroSystem

Logging enabled in ZeroSystem.

ZeroSystem

A paragraph was added to the class description.

Optimization

DenseLP

DenseLP now throws exceptions for infeasible problems.

MinUncon

Fixed resource name so that warnings are printed.

NonlinLeastSquares

Added warnings for the 4 error types.

QuadraticProgramming

A new exception was added to handle inconsistent system exceptions.

Imsl.Finance

Finance

BasisPart.GetDaysInYear

Deprecated for this release and replaced by a new BasisPart.GetDaysInYear.

Bond

Added Monthly and BiMonthly options.

Bond.Accrint

Corrected accrued interest calculation.

Bond.Yearfrac

Corrected problem where the method gave a wrong result for input days in a leap year.

DayCountBasis.GetDaysInYear

Deprecated for this release and replaced by a new DayCountBasis.GetDaysInYear.

Finance.Xirr

Modified to return better results.

Imsl.Stat

Basic Statistics

NormOneSample

Increased accuracy of p-values.

NormOneSample.ConfidenceMean

Clarified documentation.

NormTwoSample

Increased accuracy of p-values.

PartialCovariances

Increased accuracy of p-values.

TableTwoWay.GetFrequencyTableUsingClassmarks

Corrected description in documentation.

TableTwoWay.GetFrequencyTableUsingCutpoints

Corrected description in documentation.

Regression

LinearRegression.GetCoefficients

Replaced the SingularMatrixException exception by a LinearRegression.NotFullRank warning.

NonlinearRegression

Added missing NonlinearRegression resource entries.

Added warnings for the 4 error types.

RegressorsForGLM

Increased accuracy of p-values.

SelectionRegression

Modified to correctly account for missing values.

Increased accuracy of p-values.

StepwiseRegression

Increased accuracy of p-values.

Analysis of Variance

ANOVA

Increased accuracy of p-values.

ANOVA.GetDunnSidak

This method has been deprecated and replced by ANOVA.GetConfidenceInterval.

ANOVAFactorial

Increased accuracy of p-values.

Categorical and Discrete Data Analysis

CategoricalGenLinModel

Increased accuracy of p-values.

Removed confusing statement from documentation.

Added a test for rank deficiency.

CategoricalGenLinModel.LowerEndpointColumn

Corrected implementation.

ContingencyTable

Increased accuracy of p-values.

Tests of Goodness of Fit

ChiSquaredTest

Increased accuracy of p-values.

Expanded documentation of nParameters and property DegreesOfFreedom.

ChiSquaredTest

Overloaded the Update method to make the frequencies optional.

NormalityTest.LillieforsTest

Corrected the way ties are identified.

Time Series and Forecasting

ARMA

Resolved seg fault and flat forecast issue.

ARMAEstimateMissing

MaxIterations setting now applies to LeastSquares estimation. A TooManyIterationsException is thrown rather than rethrowing ARMA exceptions TooManyIterationsExceptions and TooManyFcnEvaluations which may occur during the LeastSquares estimation

Multivariate Analysis

ChiSquaredTest

Increased accuracy of p-values.

DiscriminantAnalysis

Increased accuracy of p-values.

DiscriminantAnalysis.NRowsMissing

This property has been deprecated. It has been replaced by the property NumberOfRowsMissing.

DiscriminantAnalysis.Update

Some Update methods have been deprecated. Each has been replaced by an equivalent Update method.

FactorAnalysis

Changed NoDegreesOfFreedomException exception to a warning.

Increased accuracy of p-values.

Probability Distribution Functions and Inverses

The Cdf class has been divided into the three classes Cdf, InvCdf, and Pdf. Therefore the methods which have been moved from the Cdf class have been deprecated and replaced by the functionally equivalent versions in the new class. Below is a list of the methods which have been deprecated followed by the functionally equivalent version which has been added.

Cdf.BinomialProb	Pdf.Binomial
Cdf.PoissonProb	Pdf.Poisson
Cdf.BetaProb	Pdf.Beta
Cdf.FProb	Pdf.F
Cdf.HypergeometricProb	Pdf.Hypergeometric
Cdf.GammaProb	Pdf.Gamma
Cdf.ExponentialProb	Pdf.Exponential
Cdf.ChiProb	Pdf.Chi
Cdf.WeibullProb	Pdf.Weibull
Cdf.LognormalProb	Pdf.Lognormal
Cdf.ExtremevalueProb	Pdf.Extremevalue
Cdf.RayleighProb	Pdf.Rayleigh
Cdf.DiscreteUniformProb	Pdf.DiscreteUniform
Cdf.GeometricProb	Pdf.Geometric
Cdf.InverseBeta	InvCdf.Beta
Cdf.InverseF	InvCdf.F
Cdf.InverseGamma	InvCdf.Gamma
Cdf.InverseNormal	InvCdf.Normal
Cdf.InverseChi	InvCdf.Chi
Cdf.InverseNoncentralchi	InvCdf.Noncentralchi
Cdf.InverseStudentsT	InvCdf.StudentsT
Cdf.InverseNoncentralstudentsT	InvCdf.NoncentralstudentsT
Cdf.InverseWeibull	InvCdf.Weibull
Cdf.InverseLogNormal	InvCdf.LogNormal
Cdf.InverseExponential	InvCdf.Exponential
Cdf.InverseExtremeValue	InvCdf.ExtremeValue
Cdf.InverseRayleigh	InvCdf.Rayleigh
Cdf.InverseUniform	InvCdf.Uniform
Cdf.InverseDiscreteUniform	InvCdf.DiscreteUniform
Cdf.InverseGeometric	InvCdf.Geometric

Cdf.Chi

Expanded the range of allowable arguments for degrees of freedom.

Cdf.F

Improved lower tail accuracy.

Cdf.Chi

Improved lower tail accuracy.

Cdf.Noncentralchi

Improved lower tail accuracy.

InvCdf.F

Corrected results for large denominator df.

Pdf.Gamma

Corrected calculation of gamma pdf.

Random Number Generation

General

Added Usage Notes.

Random.NextNormal

Eliminated cases where +infinity/-infinity is returned.

Imsl.Io

Imsl.Datamining.Neural

Neural Nets

ScaleFilter

Corrected possible divide by zero.

Imsl.Chart

Chart 2D

AxisLabel.Paint

Added checks to avoid printing labels beyond the extent of the chart window.

Contour

Modified logic where negative values for Max X and Y data caused distortions.

Colormap

Renamed the WHITE_BLUE_LINEAR to WB_LINEAR and documented as the reverse of BW_LINEAR. The name WHITE_BLUE_LINEAR no long exists since BLUE_WHITE should be used instead.

HeatMap

Modified logic where negative values for Max X and Y data caused distortions.