| Package | Description |
|---|---|
| gov.sandia.cognition.learning.algorithm.minimization |
Provides minimization algorithms.
|
| gov.sandia.cognition.learning.algorithm.minimization.line |
Provides line (scalar) minimization algorithms.
|
| gov.sandia.cognition.learning.algorithm.minimization.matrix |
Provides matrix solving algorithms.
|
| gov.sandia.cognition.learning.algorithm.regression |
Provides regression algorithms, such as Linear Regression.
|
| gov.sandia.cognition.statistics.method |
Provides algorithms for evaluating statistical data and conducting statistical inference, particularly frequentist methods.
|
| Modifier and Type | Class and Description |
|---|---|
class |
AbstractAnytimeFunctionMinimizer<InputType,OutputType,EvaluatorType extends Evaluator<? super InputType,? extends OutputType>>
A partial implementation of a minimization algorithm that is iterative,
stoppable, and approximate.
|
class |
FunctionMinimizerBFGS
Implementation of the Broyden-Fletcher-Goldfarb-Shanno (BFGS) Quasi-Newton
nonlinear minimization algorithm.
|
class |
FunctionMinimizerConjugateGradient
Conjugate gradient method is a class of algorithms for finding the
unconstrained local minimum of a nonlinear function.
|
class |
FunctionMinimizerDFP
Implementation of the Davidon-Fletcher-Powell (DFP) formula for a
Quasi-Newton minimization update.
|
class |
FunctionMinimizerDirectionSetPowell
Implementation of the derivative-free unconstrained nonlinear direction-set
minimization algorithm called "Powell's Method" by Numerical Recipes.
|
class |
FunctionMinimizerFletcherReeves
This is an implementation of the Fletcher-Reeves conjugate gradient
minimization procedure.
|
class |
FunctionMinimizerGradientDescent
This is an implementation of the classic Gradient Descent algorithm, also
known as Steepest Descent, Backpropagation (for neural nets), or Hill
Climbing.
|
class |
FunctionMinimizerLiuStorey
This is an implementation of the Liu-Storey conjugate gradient
minimization procedure.
|
class |
FunctionMinimizerNelderMead
Implementation of the Downhill Simplex minimization algorithm, also known as
the Nelder-Mead method.
|
class |
FunctionMinimizerPolakRibiere
This is an implementation of the Polack-Ribiere conjugate gradient
minimization procedure.
|
class |
FunctionMinimizerQuasiNewton
This is an abstract implementation of the Quasi-Newton minimization method,
sometimes called "Variable-Metric methods."
This family of minimization algorithms uses first-order gradient information
to find a locally minimum to a scalar function.
|
| Modifier and Type | Interface and Description |
|---|---|
interface |
LineMinimizer<EvaluatorType extends Evaluator<java.lang.Double,java.lang.Double>>
Defines the functionality of a line-minimization algorithm, often called a
"line search" algorithm.
|
| Modifier and Type | Class and Description |
|---|---|
class |
AbstractAnytimeLineMinimizer<EvaluatorType extends Evaluator<java.lang.Double,java.lang.Double>>
Partial AnytimeAlgorithm implementation of a LineMinimizer.
|
class |
LineMinimizerBacktracking
Implementation of the backtracking line-minimization algorithm.
|
class |
LineMinimizerDerivativeBased
This is an implementation of a line-minimization algorithm proposed by
Fletcher that makes extensive use of first-order derivative information.
|
class |
LineMinimizerDerivativeFree
This is an implementation of a LineMinimizer that does not require
derivative information.
|
| Modifier and Type | Class and Description |
|---|---|
class |
ConjugateGradientMatrixSolver
Implements a matrix solver using Conjugate Gradient.
|
class |
ConjugateGradientWithPreconditionerMatrixSolver
Implements a matrix solver using Conjugate Gradient with a preconditioner.
|
class |
IterativeMatrixSolver<Operator extends MatrixVectorMultiplier>
Base class for all iterative matrix solvers that takes care of most of the
basic iterative logic and the function minimizer interface.
|
class |
OverconstrainedConjugateGradientMatrixMinimizer
Implements a overconstrained conjugate gradient matrix optimizer.
|
class |
SteepestDescentMatrixSolver
Implements a basic Steepest Descent iterative solver for linear systems of
equations.
|
| Modifier and Type | Field and Description |
|---|---|
static FunctionMinimizer<Vector,java.lang.Double,Evaluator<? super Vector,java.lang.Double>> |
ParameterDerivativeFreeCostMinimizer.DEFAULT_FUNCTION_MINIMIZER
Default function minimizer, FunctionMinimizerDirectionSetPowell
|
static FunctionMinimizer<Vector,java.lang.Double,DifferentiableEvaluator<? super Vector,java.lang.Double,Vector>> |
ParameterDifferentiableCostMinimizer.DEFAULT_FUNCTION_MINIMIZER
Default function minimizer,
FunctionMinimizerBFGS with LineMinimizerBacktracking
|
| Constructor and Description |
|---|
AbstractMinimizerBasedParameterCostMinimizer(FunctionMinimizer<Vector,java.lang.Double,? super EvaluatorType> algorithm)
Creates a new instance of AbstractMinimizerBasedParameterCostMinimizer
|
AbstractMinimizerBasedParameterCostMinimizer(FunctionMinimizer<Vector,java.lang.Double,? super EvaluatorType> algorithm,
SupervisedCostFunction<Vector,Vector> costFunction)
Creates a new instance of AbstractMinimizerBasedParameterCostMinimizer
|
ParameterDerivativeFreeCostMinimizer(FunctionMinimizer<Vector,java.lang.Double,? super DifferentiableEvaluator<Vector,java.lang.Double,Vector>> minimizer)
Creates a new instance of ParameterDerivativeFreeCostMinimizer
|
ParameterDifferentiableCostMinimizer(FunctionMinimizer<Vector,java.lang.Double,? super DifferentiableEvaluator<Vector,java.lang.Double,Vector>> minimizer)
Creates a new instance of ParameterDerivativeFreeCostMinimizer
|
| Constructor and Description |
|---|
DistributionParameterEstimator(DistributionType distribution,
CostFunction<? super DistributionType,java.util.Collection<? extends DataType>> costFunction,
FunctionMinimizer<Vector,java.lang.Double,? super DistributionParameterEstimator.DistributionWrapper> algorithm)
Creates a new instance of DistributionParameterEstimator
|