close Warning: Can't synchronize with repository "(default)" (/var/svn/tolp does not appear to be a Subversion repository.). Look in the Trac log for more information.

  1. NameBlock
    1. Función SetToNameBlock
    2. Función StdLib::ARMAProcess::Eval.Almagro
    3. Función StdLib::ARMAProcess::FastCholeskiCovFactor
    4. Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Lat.Homog
    5. Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs
    6. Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.Data.Info
    7. Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.Input
    8. Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.Output
    9. Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.Serie.Info
    10. Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.TransFun
    11. Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.Vector.Info
    12. Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Pri.Homog
    13. Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::CreateSeriesHandler
    14. Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::LoadLatNode
    15. Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::LoadModelDef
    16. Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::LoadNodeConstraints
    17. Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::LoadObsNode
    18. Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::LoadPriNode
    19. Función …
    20. Función …
    21. Función StdLib::BysMcmc::Bsr::Gibbs::ArimaBlock
    22. Función StdLib::BysMcmc::Bsr::Gibbs::BasicMaster
    23. Función StdLib::BysMcmc::Bsr::Gibbs::BsrAsBlock
    24. Función StdLib::BysMcmc::Bsr::Gibbs::DeltaTransfer
    25. Función StdLib::BysMcmc::Bsr::Gibbs::InputMissingBlock
    26. Función StdLib::BysMcmc::Bsr::Gibbs::NoNotifier
    27. Función StdLib::BysMcmc::Bsr::Gibbs::NonLinBlock
    28. Función StdLib::BysMcmc::Bsr::Gibbs::NonLinMaster
    29. Función StdLib::BysMcmc::Bsr::Gibbs::OutputMissingBlock
    30. Función StdLib::BysMcmc::Bsr::Gibbs::ProbitFilter
    31. Función StdLib::BysMcmc::Bsr::Gibbs::SigmaBlock
    32. Función StdLib::BysMcmc::Bsr::Gibbs::StdLinearBlock
    33. Función StdLib::BysMcmc::Bsr::Import::Constraints
    34. Función StdLib::BysMcmc::Bsr::Import::Explicit.Constraints
    35. Función StdLib::BysMcmc::Bsr::Import::Generic.Constraints
    36. Función StdLib::BysMcmc::Bsr::Import::Order.Relations
    37. Función StdLib::BysMcmc::Bsr::Import::Unconstrained
    38. Función StdLib::BysMcmc::Bsr::OneNode::EstimProbit
    39. Función StdLib::BysMcmc::BuildCycler
    40. Función StdLib::BysMcmc::BuildFullConfig
    41. Función StdLib::BysMcmc::DefineBlock
    42. Función StdLib::BysMcmc::Get.Recover
    43. Función StdLib::DBConnect::Create
    44. Función StdLib::SolvePrecondSym
    45. Función StdLib::SolvePrecondUnsym
    46. Función StdLib::Timer::Start

NameBlock

Funciones que devuelven NameBlock

Función SetToNameBlock

Función StdLib::ARMAProcess::Eval.Almagro

  • Declaración: 
    NameBlock StdLib::ARMAProcess::Eval.Almagro(Polyn ar, Polyn ma, VMatrix z_, Real sigma)
  • Descripción: 
    Given an ARMA process ar(B)*z[t] = ma(B)*a[t] builds these methods:
 Draw.U_cond_Z: generate random initial values conditioned to noise
 LogLH.Z_cond_U: conditional likelihood of noise conditioned to given initial values
.
  • Lenguaje:TOL
  • Fuente : ​stdlib/tol/math/stat/models/bayesian/arima/_arma_process.tol

Función StdLib::ARMAProcess::FastCholeskiCovFactor

Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Lat.Homog

Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs

Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.Data.Info

Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.Input

Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.Output

Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.Serie.Info

Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.TransFun

Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Obs.Vector.Info

Función StdLib::BysMcmc::Bsr::DynHlm::BuildNode::Pri.Homog

Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::CreateSeriesHandler

Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::LoadLatNode

Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::LoadModelDef

Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::LoadNodeConstraints

Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::LoadObsNode

Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::LoadPriNode

Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::ModSes.Lat.Child.Param.Def

Función StdLib::BysMcmc::Bsr::DynHlm::DBApi::ModSes.Lat.Father.Param.Def

Función StdLib::BysMcmc::Bsr::Gibbs::ArimaBlock

Función StdLib::BysMcmc::Bsr::Gibbs::BasicMaster

Función StdLib::BysMcmc::Bsr::Gibbs::BsrAsBlock

Función StdLib::BysMcmc::Bsr::Gibbs::DeltaTransfer

Función StdLib::BysMcmc::Bsr::Gibbs::InputMissingBlock

Función StdLib::BysMcmc::Bsr::Gibbs::NoNotifier

Función StdLib::BysMcmc::Bsr::Gibbs::NonLinBlock

Función StdLib::BysMcmc::Bsr::Gibbs::NonLinMaster

  • Declaración: 
    NameBlock StdLib::BysMcmc::Bsr::Gibbs::NonLinMaster(Set modelDef, Set nonLinFilter, NameBlock config)
  • Descripción: 
    Builds a NameBlock that can draw a Gibbs sample of a Bayesian Sparse Regression model with non linear blocks by passing it to method BysMcmc::BuildCycler
  • Lenguaje:TOL
  • Fuente : ​stdlib/tol/math/stat/models/bayesian/bysMcmc/bsr/gibbs/_nonLinMaster.tol

Función StdLib::BysMcmc::Bsr::Gibbs::OutputMissingBlock

Función StdLib::BysMcmc::Bsr::Gibbs::ProbitFilter

Función StdLib::BysMcmc::Bsr::Gibbs::SigmaBlock

Función StdLib::BysMcmc::Bsr::Gibbs::StdLinearBlock

Función StdLib::BysMcmc::Bsr::Import::Constraints

Función StdLib::BysMcmc::Bsr::Import::Explicit.Constraints

Función StdLib::BysMcmc::Bsr::Import::Generic.Constraints

Función StdLib::BysMcmc::Bsr::Import::Order.Relations

Función StdLib::BysMcmc::Bsr::Import::Unconstrained

Función StdLib::BysMcmc::Bsr::OneNode::EstimProbit

  • Declaración: 
    NameBlock StdLib::BysMcmc::Bsr::OneNode::EstimProbit(NameBlock data, NameBlock config_)
  • Descripción: 
    Estimates a probit model with just one node, usually observational and dense.Model is defined as a BSR basic and an non linear ProbitFilter.BSR will be built from simplified data argument matching the same API used by BysMcmc::Bsr::Gibbs::EstimOneNode, fixing _.sigma to 1NameBlock data = 
[[
//Mandatory members 
  Set  _.docInfo         //BSR.DocInfo 
  Text _.segmentName;    //Node name
  Set _.linearParamInfo; //Set of Bsr.OneNode.LinearParamInfo
  Anything _.Y;          //Output data (Matrix or VMatrix)
  Anything _.X;          //Input data (Matrix or VMatrix)
  Real _.sigma = 1;      //Sigma value or ? to simulate it
//Optional members 
  Set _.orderRelation;   //Set of Bsr.OrderRelation.Info
  Set _.arima;           //Set of ARIMAStruct
  Set _.timeInfo;        //Set of BSR.NoiseTimeInfo
]];
  • Lenguaje:TOL
  • Fuente : ​stdlib/tol/math/stat/models/bayesian/bysMcmc/bsr/_oneNode.tol

Función StdLib::BysMcmc::BuildCycler

  • Declaración: 
    NameBlock StdLib::BysMcmc::BuildCycler(NameBlock modelSampler, NameBlock config, NameBlock notifier)
  • Descripción: 
    Builds a NameBlock that is able to generate a Markov Chain of a model by cycling calls of a given individual drawer and gives a set of tools to make Bayesian Inference about it. 
NameBlock modelSampler is a handler of the model to be simulated and must have at least these public methods: 
 Text get.name     (Real unused); //Model name
 Text get.session  (Real unused); //Session tag name
 Text get.path     (Real unused); //Path to store Markov Chain as BBM
 Set  get.colNames (Real unused); //Names of Markov Chain variables
 Matrix draw       (Real numSim); //Draws a simulation of Markov Chain
  • Lenguaje:TOL
  • Fuente : ​stdlib/tol/math/stat/models/bayesian/bysMcmc/_build.tol

Función StdLib::BysMcmc::BuildFullConfig

Función StdLib::BysMcmc::DefineBlock

Función StdLib::BysMcmc::Get.Recover

Función StdLib::DBConnect::Create

  • Declaración: 
    NameBlock StdLib::DBConnect::Create(Text alias, Text user, Text password, Text driver, Text defaultDataBase, Text server, Text purpose)
  • Descripción: 
  • Lenguaje:TOL

  • Fuente : ​stdlib/tol/data/db/_db_connect.tol 

Función StdLib::SolvePrecondSym


  • Declaración:

    NameBlock StdLib::SolvePrecondSym(VMatrix M, Real do.normalize)

  • Descripción:

    Soves a symmetric linear system, that could be large, sparse and ill-conditioned, in a robust and fast way by applying diagonal additive preconditioner of previously normalized system foilowing these steps:
1. Original symmetric system M*z = y is previously normalized by non null elements of root squares of diagonal elements of M that will called D:
  S*u = v where S=Di*M*Di; Di=D^-1; u=D*x; v=Di*y
Then
  Di*M*Di*D*x = Di*y <=> M*x=y 
2. A digonal preconditioner (S+c*I)^-1 will be built to be (S+c*I)^-1*S close to identity and number c is found to be the lesser than S+c*I admits Choleski decomposition S+c*I = L*L'. 
3. Preconditioned system is solved by Lanczos iterative method with partial reorthogonalization MinimumResidualsSolve. Preconditioner is stored as a Code operator using just sparse matrix L to avoid store ((S+c*I)^-1)*S that is not neccessary so sparse as S nor L.
4. Solution is denormalized to solve original problem.
Normalization is an optional task that is specified by argument do.normalize
If M is numerically definite positive (c=0) then CholeskiSolve will be used directly if argument forze.lanczos of Solve method is true.
Using instructions to solve M*x=y :
  //Builds preconditioning handler with normalization
  NameBlock SPDP = SymPosDefPrecond(H, True); 
  //Finds optimal preconditioner
  Real SPDP::ScaleOptimize(1.E-7); 
  //Solves the system
  NameBlock solution = SPDP::Solve(y,1.E-10,True); 
  //Takes the solution matrix
  VMatrix x = solution::_.x;



  • Lenguaje:TOL

  • Fuente : ​stdlib/tol/math/linalg/_solve_precond_sym.tol 

Función StdLib::SolvePrecondUnsym


  • Declaración:

    NameBlock StdLib::SolvePrecondUnsym(VMatrix M, Real do.normalize)

  • Descripción:

    Soves an unsymmetric overdetermined linear system, that could be large, sparse and ill-conditioned, in a robust and fast way by applying diagonal additive preconditioner of previously normalized system foilowing these steps:
1. Original unsymmetric system M*z = y is previously normalized by non null elements of root squares of diagonal elements of M'M that will called D:
  H*u = y where H=M*Di; Di=D^-1; u=D*x;
2. A digonal expansion preconditioner Hc=(H<<c*I)^-1 will be built to be 
  Hc'*Hc=(S+c*I)^-1*S; where S=H'*H 
close to identity and number c is found to be the lesser than H<<c*I admits Choleski decomposition S+c*I = Lc*Lc'. 
3. Preconditioned system is solved by Lanczos iterative method with partial reorthogonalization MinimumResidualsSolve. Preconditioner is stored as a Code operator using just sparse matrix Lc to avoid store ((S+c*I)^-1)*S that is not neccessary so sparse as S nor L.
4. Solution is denormalized to solve original problem.
Using instructions to solve M*x=y :
  //Builds preconditioning handler
  NameBlock SPDP = SymPosDefPrecond(H); 
  //Finds optimal preconditioner
  Real SPDP::ScaleOptimize(1.E-7); 
  //Solves the system
  NameBlock solution = SPDP::Solve(y,1.E-10); 
  //Takes the solution matrix
  VMatrix x = solution::_.x;



  • Lenguaje:TOL

  • Fuente : ​stdlib/tol/math/linalg/_solve_precond_unsym.tol 

Función StdLib::Timer::Start



          
          

            Last modified 16 years ago
            Last modified on Feb 27, 2009, 5:50:05 PM