Sampling methods |
Candidate densities |
Marginal likelihood computation methods |
MC_MH | Metropolis Hastings, see e.g. [Chib and Greenberg (1995)] |
MC_IS | Importance Sampling [Kloek and Van Dijk (1978)] |
MC_APS | Adaptive Polar Sampling, [Bauwens et al. (2004)] |
MC_APIS | Adaptive Polar Importance Sampling [Bauwens et al. (2004)] |
MC_GG | Griddy Gibbs Sampling [Ritter and Tanner (1992)] |
MC_CNORM | Normal candidate density (with location equal to preset value, scale a multiplication of last scale) |
MC_CSTUD | Student-t candidate density with df degrees of freedom, (with location equal to preset value, scale a multiplication of last scale) |
MC_CRW | Random walk candidate density with covariance matrix proportional to the last estimate of the covariance matrix |
MC_CUSER | User specified candidate density |
MC_MLKERN | Use a kernel approximation to the posterior density, to compute the marginal likelihood as the difference between the log-density and log-posterior kernel at a specific location |
MC_MLLP | Use a LaPlace approximation to the posterior density, to compute the marginal likelihood as the difference between the log-density and log-posterior kernel at a specific location |
MC_MLHM | Compute the likelihood over all sampled data points, and compute the marginal likelihood as the harmonic mean of these densities. |
MC_MLGG | Use a series of Gibbs chains, and derive the the marginal likelihood from the conditional densities. |
MC_MLINT | Use a pure multidimensional numerical integration to compute the the marginal likelihood. |
MC_MLPR | Sample from the prior and average over the corresponding values of the likelihood to compute the marginal likelihood. |
MC2Pack::GetDraws(const amTheta, const avW)
MC2Pack::GetLocationScale()
MC2Pack::GetPackageName()
MC2Pack::GetPackageVersion()
MC2Pack::GetParNames()
MC2Pack::GetRNE()
|
MC2Pack::GetTaper()
MC2Pack::SetCandidate(const iCand, const xCand);
fnRndCand(const iR, const vTheta, const vMu, const mCS2, const amCand, const avLnDens)with inputs
MC2Pack::SetConditional(const vInd, const fnRanCondGG) MC2Pack::SetConditional(const vInd, const fnRanCondGG, const fnDensCondGG)
MC2Pack::SetDebug(const iDebug)
MC2Pack::SetEps(const dEpsAbs, const dEpsRel)
MC2Pack::SetGraphs(const bShowGraphs)
MC2Pack::SetInfo(const iInfoRep)
MC2Pack::SetIntegration(const iAdaptive, ...) MC2Pack::SetIntegration(const iAdaptive, const iBounds, const bSMax, const bSeparate)
MC2Pack::SetLikelihood(const fnLikelihood)
MC2Pack::SetLimits(const mLUBounds) MC2Pack::SetLimits(const mLUBounds, const mBC) MC2Pack::SetLimits(const mLUBounds, const mBC, const bChangeLU)
MC2Pack::SetLocationScale(const avMu, const amS2, const iN, const bOptimize, ...) MC2Pack::SetLocationScale(const avMu, const amS2, const iN, const bOptimize, const asNames)
MC2Pack::SetMahalanobisFraction(const dFrac)
MC2Pack::SetMethod(const iMethod)
MC2Pack::SetOutput(const sResbase) MC2Pack::SetOutput(const sResbase, const bRewrite)
MC2Pack::SetParNames(const asNames)
MC2Pack::SetPosterior(const fnPosterior)
MC2Pack::SetPrior(const fnPrior)
MC2Pack::SetRanPrior(const fnRanPrior)
mU= fnRanPrior(iR);
MC2Pack::SetSample(const vRep, ...)
MC2Pack::SetTaper(const vTaper)
MC2Pack::Cusum(const bAdapt)
MC2Pack::CumPlot(const viCum)
Marglik(const iMethod, ...)The actual inputs depend on the method which is chosen for computing the marginal likelihood. Possible inputs are:
Method | Arguments | Functions used |
MC_MLKERN | vMu | Posterior |
MC_MLLP | vMu, mS2 | Posterior |
MC_MLHM | dFrac | Likelihood |
MC_MLGG | vMu, nRep | Posterior, Conditionals (if available) |
MC_MLINT | nInt | Posterior |
MC_MLPR | nRep | Likelihood, RanPrior |
MC_MLIS | vMu, mS2, nRep | Prior, Likelihood |
MC2Pack::Simulate()