Package gov.sandia.cognition.statistics.bayesian

Provides algorithms for computing Bayesian estimates of parameters.

See:
          Description

Interface Summary
BayesianEstimator<ObservationType,ParameterType,PosteriorType extends Distribution<? extends ParameterType>> A type of estimation procedure based on Bayes's rule, which allows us to estimate the uncertainty of parameters given a set of observations that we are given.
BayesianEstimatorPredictor<ObservationType,ParameterType,PosteriorType extends Distribution<? extends ParameterType>> A BayesianEstimator that can also compute the predictive distribution of new data given the posterior.
BayesianParameter<ParameterType,ConditionalType extends Distribution<?>,PriorType extends Distribution<ParameterType>> A parameter from a Distribution that has an assumed Distribution of values.
BayesianRegression<OutputType,PosteriorType extends Distribution<? extends Vector>> A type of regression algorithm maps a Vector space, and the weights of this Vector space are represented as a posterior distribution given the observed InputOutputPairs.
DirichletProcessMixtureModel.Updater<ObservationType> Updater for the DPMM
ImportanceSampling.Updater<ObservationType,ParameterType> Updater for ImportanceSampling
MarkovChainMonteCarlo<ObservationType,ParameterType> Defines the functionality of a Markov chain Monte Carlo algorithm.
MetropolisHastingsAlgorithm.Updater<ObservationType,ParameterType> Creates proposals for the MCMC steps.
ParticleFilter<ObservationType,ParameterType> A particle filter aims to estimate a sequence of hidden parameters based on observed data using point-mass estimates of the posterior distribution.
ParticleFilter.Updater<ObservationType,ParameterType> Updates the particles.
RecursiveBayesianEstimator<ObservationType,ParameterType,BeliefType extends Distribution<ParameterType>> A recursive Bayesian estimator is an estimation method that uses the previous belief of the system parameter and a single observation to refine the estimate of the system parameter.
RejectionSampling.Updater<ObservationType,ParameterType> Updater for ImportanceSampling
 

Class Summary
AbstractBayesianParameter<ParameterType,ConditionalType extends ClosedFormDistribution<?>,PriorType extends Distribution<ParameterType>> Partial implementation of BayesianParameter
AbstractKalmanFilter Contains fields useful to both Kalman filters and extended Kalman filters.
AbstractMarkovChainMonteCarlo<ObservationType,ParameterType> Partial abstract implementation of MarkovChainMonteCarlo.
AbstractParticleFilter<ObservationType,ParameterType> Partial abstract implementation of ParticleFilter.
AdaptiveRejectionSampling Samples form a univariate distribution using the method of adaptive rejection sampling, which is a very efficient method that iteratively improves the rejection and acceptance envelopes in response to additional points.
AdaptiveRejectionSampling.LineSegment A line that has a minimum and maximum support (x-axis) value.
AdaptiveRejectionSampling.LogEvaluator<EvaluatorType extends Evaluator<Double,Double>> Wraps an Evaluator and takes the natural logarithm of the evaluate method
AdaptiveRejectionSampling.PDFLogEvaluator Wraps a PDF so that it returns the logEvaluate method.
AdaptiveRejectionSampling.Point An InputOutputPair that has a natural ordering according to their input (x-axis) values.
BayesianCredibleInterval A Bayesian credible interval defines a bound that a scalar parameter is within the given interval.
BayesianLinearRegression Computes a Bayesian linear estimator for a given feature function and a set of observed data.
BayesianLinearRegression.IncrementalEstimator Incremental estimator for BayesianLinearRegression
BayesianRobustLinearRegression Computes a Bayesian linear estimator for a given feature function given a set of InputOutputPair observed values.
BayesianRobustLinearRegression.IncrementalEstimator Incremental estimator for BayesianRobustLinearRegression
BayesianUtil Contains generally useful utilities for Bayesian statistics.
DefaultBayesianParameter<ParameterType,ConditionalType extends ClosedFormDistribution<?>,PriorType extends Distribution<ParameterType>> Default implementation of BayesianParameter using reflection.
DirichletProcessMixtureModel<ObservationType> An implementation of Dirichlet Process clustering, which estimates the number of clusters and the centroids of the clusters from a set of data.
DirichletProcessMixtureModel.DPMMCluster<ObservationType> Cluster for a step in the DPMM
DirichletProcessMixtureModel.DPMMLogConditional Container for the log conditional likelihood
DirichletProcessMixtureModel.MultivariateMeanCovarianceUpdater Updater that creates specified clusters with distinct means and covariances
DirichletProcessMixtureModel.MultivariateMeanUpdater Updater that creates specified clusters with identical covariances
DirichletProcessMixtureModel.Sample<ObservationType> A sample from the Dirichlet Process Mixture Model.
ExtendedKalmanFilter Implements the Extended Kalman Filter (EKF), which is an extension of the Kalman filter that allows nonlinear motion and observation models.
GaussianProcessRegression<InputType> Gaussian Process Regression, is also known as Kriging, is a nonparametric method to interpolate and extrapolate using Bayesian regression, where the expressiveness of the estimator can grow with the data.
ImportanceSampling<ObservationType,ParameterType> Importance sampling is a Monte Carlo inference technique where we sample from an easy distribution over the hidden variables (parameters) and then weight the result by the ratio of the likelihood of the parameters given the evidence and the likelihood of generating the parameters.
ImportanceSampling.DefaultUpdater<ObservationType,ParameterType> Default ImportanceSampling Updater that uses a BayesianParameter to compute the quantities of interest.
KalmanFilter A Kalman filter estimates the state of a dynamical system corrupted with white Gaussian noise with observations that are corrupted with white Gaussian noise.
MetropolisHastingsAlgorithm<ObservationType,ParameterType> An implementation of the Metropolis-Hastings MCMC algorithm, which is the most general formulation of MCMC but can be slow.
ParallelDirichletProcessMixtureModel<ObservationType> A Parallelized version of vanilla Dirichlet Process Mixture Model learning.
ParallelDirichletProcessMixtureModel.DPMMAssignments Assignments from the DPMM
RejectionSampling<ObservationType,ParameterType> Rejection sampling is a method of inferring hidden parameters by using an easy-to-sample-from distribution (times a scale factor) that envelopes another distribution that is difficult to sample from.
RejectionSampling.DefaultUpdater<ObservationType,ParameterType> Default ImportanceSampling Updater that uses a BayesianParameter to compute the quantities of interest.
RejectionSampling.ScalarEstimator<ObservationType> Routine for estimating the minimum scalar needed to envelop the conjunctive distribution.
SamplingImportanceResamplingParticleFilter<ObservationType,ParameterType> An implementation of the standard Sampling Importance Resampling particle filter.
 

Package gov.sandia.cognition.statistics.bayesian Description

Provides algorithms for computing Bayesian estimates of parameters.

Since:
3.0
Author:
Kevin R. Dixon