Mantis - ISSUES

0000375: INCOMPATIBLE: nonlinear fitting functions were renamed

LSFitNonlinear???() functions were renamed to LSFit???() to make their names compatible with naming conventions used by Optimization package.

0000373: Pseudoinverse

real

0000371: IMPLEMENTED: nonlinear solver

New nonlinear Levenberg-Marquardt-like solver was implemented. Its essential features are:
* weak requirements on the system being solved (Jacobian non-singularity assumption is dropped)
* global convergence
* quadratic convergence in the neighborhood of solution
* same set of additional options as for optimization algorithms: reports, limit on step size, efficient restarts

0000370: IMPLEMENTED: buffered functions for optimization algorithms

Optimization algorithms have functions with common names like 'ALGONAMEResults' which were used to get algorithm result. These functions have output parameter X - array, which gets reallocated on each call (because of semantics of these functions).

This fix adds new family of functions - 'ALGONAMEResultsBuf', which takes preallocated buffer (and reallocates it, if too small). It allows to save time when algorithm is used from inner cycle of another performance critical algorithm.

0000369: INCOMPATIBLE: changes in 'nearestneighbor' unit

Backward compatible changes:
* buffered query functions were added. These functions can use preallocated buffer (which is automatically reallocated, if it is too small), which reduces number of dynamic allocations and increases overall performance
* better detection of errors in input parameters

Incompatible changes:
* functions from 'KDTreeQueryResults' family do NOT return result size K anymore. This value is already returned by functions from 'KDTreeQuery' family, so there are no reasons to duplicate it.

0000368: More support for IEEE special quantities

Check for NANs/INFs in the input data
* linear algebra/interpolation - check before starting
* optimization (or other rcomm's) - check every time we get rcomm data
* statistics (correlation - with respect to missing values)
* generally, check all parameters

0000353: IMPLEMENTED: improved support for IEEE special quantities

Several units now support detection of IEEE special quantities in the input data. These units are polint, ratint, matdet, matinv.

0000358: INCOMPATIBLE: support for LDLT factorization is dropped from ALGLIB 3.0

LDLT factorization is dropped from ALGLIB 3.0

0000367: FIXED: documentation bug

The description of the complex number routines in visual basic defines the addition routines as multiplications.

"Public Function C_Add(Z1 As Complex Z2 As Complex):Complex
Public Function C_AddR(Z1 As Complex R As Double):Complex
Calculate Z1*Z2 or Z1*R."

0000348: IMPLEMENTED: RestartFrom() functionality for optimizers

All optimizers can be restarted from new point without allocating new optimizer object. It allows to significantly increase performance when solving many sequential optimization problems with same sizes.

0000366: New spline 2D operations

* list of local minimums
* global minimum

0000365: New 1D spline operations

* list of local extremums
* global extremum
* list of zeros

0000364: Multiple 'prepare' calls must have no allocation penalty

todo

0000363: FIXED: bug in UnequalVarianceTTest

Degrees of freedom for t-statistic were incorrectly calculated.

Thanks to Gerhard Vogt for pointing out this bug.

0000357: User-friendly BLAS

* operations with matrices with much more simpler interface than that of ABLAS unit: add, sub, FMA
* high-level operations: sin/cos/exp/pow/...

0000360: INCOMPATIBLE: renamed several special functions

Due to name conflicts with C standard library following names were changed:
* Gamma => GammaFunction
* Erf => ErrorFunction
* ErfC => ErrorFunctionC

0000359: New serialization framework

* generate serialization code for structs with 'object', 'optimizer' or other object-like interface modifiers.
* preserve backward compatible interface

0000282: Symmetric random matrix must be exactly symmetric

it must be checked in unit test

0000355: Non-LM nonlinear fitting

Use L-BFGS or CG, if user wants us to.
May be useful when we work with high dimensional fitting problems.

0000354: Fitting functions for important special cases

* fitting by sphere
* total least squares straight line
* fitting by Gaussian (with offset, without offset, with prescribed coefficients)

0000352: Classification improvements

1. document bdss.ap functions
2. extended linear classification subroutines:
* binary classification, minimization of RMS or CV
* binary classification, minimal cost
* applly them to LDA

0000343: Assertion on interpolation problems with non-distinct points

* Assert() when there are points with equal abscissas
* we should go beyond simple check for equality because of problems with abscissas like [0, 1E-300, 1, 2, 3, ...]. The best thing found so far is to map all points from [min,max] to [1,2] and to check for equality AFTER mapping is done.

0000339: Make C# program ref/out-correct

* "AlgoPascal-out" parameters should be "C#-out" too
* "AlgoPascal-const" should be C#-no-modifiers (currently we pass constant array with "ref" although it always passed by value)

0000333: ap::complex power

It is useful to implement pow method for ap::complex class, like:

complex& operator^=(const double& p) {
double r = pow(sqrt(x*x + y*y), p);
double phi = p * (x >= 0. ? atan(y/x) : (y >= 0. ? atan(y/x) + M_PI : atan(y/x) - M_PI));
x = r * cos(phi);
y = r * sin(phi);
return *this;
};

0000258: Special functions for spline/rational interpolation on equidistant grid.

something like polintbuildeqdist (no separate array for nodes)