Search results for "Statistics & Probability"

showing 10 items of 436 documents

Importance sampling type estimators based on approximate marginal Markov chain Monte Carlo

2020

We consider importance sampling (IS) type weighted estimators based on Markov chain Monte Carlo (MCMC) targeting an approximate marginal of the target distribution. In the context of Bayesian latent variable models, the MCMC typically operates on the hyperparameters, and the subsequent weighting may be based on IS or sequential Monte Carlo (SMC), but allows for multilevel techniques as well. The IS approach provides a natural alternative to delayed acceptance (DA) pseudo-marginal/particle MCMC, and has many advantages over DA, including a straightforward parallelisation and additional flexibility in MCMC implementation. We detail minimal conditions which ensure strong consistency of the sug…

Statistics and ProbabilityHyperparameter05 social sciencesBayesian probabilityStrong consistencyEstimatorContext (language use)Markov chain Monte Carlo01 natural sciencesStatistics::Computation010104 statistics & probabilitysymbols.namesake0502 economics and businesssymbols0101 mathematicsStatistics Probability and UncertaintyParticle filterAlgorithmImportance sampling050205 econometrics MathematicsScandinavian Journal of Statistics
researchProduct

The smallest singular value of a shifted $d$-regular random square matrix

2017

We derive a lower bound on the smallest singular value of a random d-regular matrix, that is, the adjacency matrix of a random d-regular directed graph. Specifically, let $$C_1<d< c n/\log ^2 n$$ and let $$\mathcal {M}_{n,d}$$ be the set of all $$n\times n$$ square matrices with 0 / 1 entries, such that each row and each column of every matrix in $$\mathcal {M}_{n,d}$$ has exactly d ones. Let M be a random matrix uniformly distributed on $$\mathcal {M}_{n,d}$$ . Then the smallest singular value $$s_{n} (M)$$ of M is greater than $$n^{-6}$$ with probability at least $$1-C_2\log ^2 d/\sqrt{d}$$ , where c, $$C_1$$ , and $$C_2$$ are absolute positive constants independent of any other parameter…

Statistics and ProbabilityIdentity matrixAdjacency matrices01 natural sciencesSquare matrixCombinatorics010104 statistics & probabilityMatrix (mathematics)Mathematics::Algebraic GeometryFOS: MathematicsMathematics - Combinatorics60B20 15B52 46B06 05C80Adjacency matrix0101 mathematicsCondition numberCondition numberMathematicsRandom graphsRandom graphLittlewood–Offord theorySingularity010102 general mathematicsProbability (math.PR)InvertibilityRegular graphsSingular valueSmallest singular valueAnti-concentrationSingular probabilitySparse matricesCombinatorics (math.CO)Statistics Probability and UncertaintyRandom matricesRandom matrixMathematics - ProbabilityAnalysis
researchProduct

Multiple smoothing parameters selection in additive regression quantiles

2021

We propose an iterative algorithm to select the smoothing parameters in additive quantile regression, wherein the functional forms of the covariate effects are unspecified and expressed via B-spline bases with difference penalties on the spline coefficients. The proposed algorithm relies on viewing the penalized coefficients as random effects from the symmetric Laplace distribution, and it turns out to be very efficient and particularly attractive with multiple smooth terms. Through simulations we compare our proposal with some alternative approaches, including the traditional ones based on minimization of the Schwarz Information Criterion. A real-data analysis is presented to illustrate t…

Statistics and ProbabilityIterative methodSchall algorithmexible modellingMathematicsofComputing_NUMERICALANALYSISAdditive quantile regression030229 sport sciencesP splines01 natural sciencesRegressionQuantile regression010104 statistics & probability03 medical and health sciences0302 clinical medicineP-splineStatisticsCovariatesemiparametric quantile regression0101 mathematicsStatistics Probability and UncertaintySmoothingSelection (genetic algorithm)QuantileMathematicsStatistical Modelling
researchProduct

Local bandwidth selection for kernel density estimation in a bifurcating Markov chain model

2020

International audience; We propose an adaptive estimator for the stationary distribution of a bifurcating Markov Chain onRd. Bifurcating Markov chains (BMC for short) are a class of stochastic processes indexed by regular binary trees. A kernel estimator is proposed whose bandwidths are selected by a method inspired by the works of Goldenshluger and Lepski [(2011), 'Bandwidth Selection in Kernel Density Estimation: Oracle Inequalities and Adaptive Minimax Optimality',The Annals of Statistics3: 1608-1632). Drawing inspiration from dimension jump methods for model selection, we also provide an algorithm to select the best constant in the penalty. Finally, we investigate the performance of the…

Statistics and ProbabilityKernel density estimationadaptive estimationNonparametric kernel estimation01 natural sciences010104 statistics & probability[MATH.MATH-ST]Mathematics [math]/Statistics [math.ST]0502 economics and businessbinary treesApplied mathematicsbifurcating autoregressive processes0101 mathematics[MATH]Mathematics [math]050205 econometrics MathematicsBinary treeStationary distributionMarkov chainStochastic processModel selection05 social sciencesEstimator[MATH.MATH-PR]Mathematics [math]/Probability [math.PR]Adaptive estimatorStatistics Probability and UncertaintyGoldenshluger-Lepski methodology
researchProduct

Posterior moments and quantiles for the normal location model with Laplace prior

2021

We derive explicit expressions for arbitrary moments and quantiles of the posterior distribution of the location parameter η in the normal location model with Laplace prior, and use the results to approximate the posterior distribution of sums of independent copies of η.

Statistics and ProbabilityLaplace priorsLaplace priorLocation parameterreflected generalized gamma priorSettore SECS-P/05Posterior probability0211 other engineering and technologiesSettore SECS-P/05 - Econometria02 engineering and technology01 natural sciencesCornish-Fisher approximation010104 statistics & probabilityStatistics::Methodologyposterior quantile0101 mathematicsposterior moments and cumulantsMathematicsreflected generalized gamma priors021103 operations researchLaplace transformLocation modelMathematical analysisStatistics::Computationposterior moments and cumulantCornish–Fisher approximationSettore SECS-S/01 - StatisticaNormal location modelposterior quantilesQuantileCommunications in Statistics - Theory and Methods
researchProduct

Local Asymptotic Normality for Shape and Periodicity in the Drift of a Time Inhomogeneous Diffusion

2017

We consider a one-dimensional diffusion whose drift contains a deterministic periodic signal with unknown periodicity $T$ and carrying some unknown $d$-dimensional shape parameter $\theta$. We prove Local Asymptotic Normality (LAN) jointly in $\theta$ and $T$ for the statistical experiment arising from continuous observation of this diffusion. The local scale turns out to be $n^{-1/2}$ for the shape parameter and $n^{-3/2}$ for the periodicity which generalizes known results about LAN when either $\theta$ or $T$ is assumed to be known.

Statistics and ProbabilityLocal asymptotic normalityMathematical analysisLocal scale62F12 60J60020206 networking & telecommunicationsMathematics - Statistics Theory02 engineering and technologyStatistics Theory (math.ST)01 natural sciencesShape parameterPeriodic function010104 statistics & probability0202 electrical engineering electronic engineering information engineeringFOS: Mathematics0101 mathematicsDiffusion (business)Mathematics
researchProduct

Bayesian analysis of a disability model for lung cancer survival

2016

Bayesian reasoning, survival analysis and multi-state models are used to assess survival times for Stage IV non-small-cell lung cancer patients and the evolution of the disease over time. Bayesian estimation is done using minimum informative priors for the Weibull regression survival model, leading to an automatic inferential procedure. Markov chain Monte Carlo methods have been used for approximating posterior distributions and the Bayesian information criterion has been considered for covariate selection. In particular, the posterior distribution of the transition probabilities, resulting from the multi-state model, constitutes a very interesting tool which could be useful to help oncolog…

Statistics and ProbabilityLung NeoplasmsEpidemiologyComputer scienceMatemáticasPosterior probabilityBayesian probabilityEstadísticaBiostatisticsAccelerated failure time modelsBayesian inference01 natural sciences010104 statistics & probability03 medical and health sciencesBayes' theoremsymbols.namesake0302 clinical medicineHealth Information ManagementBayesian information criterionCarcinoma Non-Small-Cell LungStatisticsPrior probabilityHumans0101 mathematicsBiología y BiomedicinaNeoplasm StagingInformáticaBayes estimatorBayes TheoremMarkov chain Monte CarloSurvival AnalysisBayesian information criterionMarkov Chains030220 oncology & carcinogenesisMinimum informative priorsymbolsMulti-state modelsRegression AnalysisWeibull distributionMonte Carlo Method
researchProduct

Ergodicity and limit theorems for degenerate diffusions with time periodic drift. Application to a stochastic Hodgkin−Huxley model

2016

We formulate simple criteria for positive Harris recurrence of strongly degenerate stochastic differential equations with smooth coefficients on a state space with certain boundary conditions. The drift depends on time and space and is periodic in the time argument. There is no time dependence in the diffusion coefficient. Control systems play a key role, and we prove a new localized version of the support theorem. Beyond existence of some Lyapunov function, we only need one attainable inner point of full weak Hoermander dimension. Our motivation comes from a stochastic Hodgkin−Huxley model for a spiking neuron including its dendritic input. This input carries some deterministic periodic si…

Statistics and ProbabilityLyapunov function010102 general mathematicsErgodicityDegenerate energy levels01 natural sciencesPeriodic function010104 statistics & probabilitysymbols.namesakeStochastic differential equationsymbolsState spaceApplied mathematicsLimit (mathematics)0101 mathematicsBrownian motionMathematicsESAIM: Probability and Statistics
researchProduct

Recursive estimation of the conditional geometric median in Hilbert spaces

2012

International audience; A recursive estimator of the conditional geometric median in Hilbert spaces is studied. It is based on a stochastic gradient algorithm whose aim is to minimize a weighted L1 criterion and is consequently well adapted for robust online estimation. The weights are controlled by a kernel function and an associated bandwidth. Almost sure convergence and L2 rates of convergence are proved under general conditions on the conditional distribution as well as the sequence of descent steps of the algorithm and the sequence of bandwidths. Asymptotic normality is also proved for the averaged version of the algorithm with an optimal rate of convergence. A simulation study confirm…

Statistics and ProbabilityMallows-Wasserstein distanceRobbins-Monroasymptotic normalityCLTcentral limit theoremAsymptotic distributionMathematics - Statistics TheoryStatistics Theory (math.ST)01 natural sciencesMallows–Wasserstein distanceonline data010104 statistics & probability[MATH.MATH-ST]Mathematics [math]/Statistics [math.ST]60F05FOS: MathematicsApplied mathematics[ MATH.MATH-ST ] Mathematics [math]/Statistics [math.ST]0101 mathematics62L20MathematicsaveragingSequential estimation010102 general mathematicsEstimatorRobbins–MonroConditional probability distribution[STAT.TH]Statistics [stat]/Statistics Theory [stat.TH]Geometric medianstochastic gradient[ STAT.TH ] Statistics [stat]/Statistics Theory [stat.TH]robust estimatorRate of convergenceConvergence of random variablesStochastic gradient.kernel regressionsequential estimationKernel regressionStatistics Probability and Uncertainty
researchProduct

Pairwise Markov properties for regression graphs

2016

With a sequence of regressions, one may generate joint probability distributions. One starts with a joint, marginal distribution of context variables having possibly a concentration graph structure and continues with an ordered sequence of conditional distributions, named regressions in joint responses. The involved random variables may be discrete, continuous or of both types. Such a generating process specifies for each response a conditioning set that contains just its regressor variables, and it leads to at least one valid ordering of all nodes in the corresponding regression graph that has three types of edge: one for undirected dependences among context variables, another for undirect…

Statistics and ProbabilityMarkov chain010102 general mathematicsMixed graphConditional probability distribution01 natural sciencesCombinatorics010104 statistics & probabilityConditional independenceJoint probability distributionMarkov property0101 mathematicsStatistics Probability and UncertaintyMarginal distributionRandom variableMathematicsStat
researchProduct