Copyright	(c) Adam Scibior 2015-2020
License	MIT
Maintainer	leonhard.markert@tweag.io
Stability	experimental
Portability	GHC
Safe Haskell	Safe-Inferred
Language	Haskell2010

Control.Monad.Bayes.Traced.Static

Description

Synopsis

data TracedT m a = TracedT {
- model :: WeightedT (DensityT m) a
- traceDist :: m (Trace a)
}
hoist :: (forall x. m x -> m x) -> TracedT m a -> TracedT m a
marginal :: Monad m => TracedT m a -> m a
mhStep :: MonadDistribution m => TracedT m a -> TracedT m a
mh :: MonadDistribution m => Int -> TracedT m a -> m [a]

Documentation

data TracedT m a Source #

A tracing monad where only a subset of random choices are traced.

The random choices that are not to be traced should be lifted from the transformed monad.

Constructors

TracedT
Fields model :: WeightedT (DensityT m) a traceDist :: m (Trace a)

Instances

Instances details

MonadTrans TracedT Source #
Instance details Defined in Control.Monad.Bayes.Traced.Static Methods lift :: Monad m => m a -> TracedT m a #
Monad m => Applicative (TracedT m) Source #
Instance details Defined in Control.Monad.Bayes.Traced.Static Methods pure :: a -> TracedT m a # (<>) :: TracedT m (a -> b) -> TracedT m a -> TracedT m b # liftA2 :: (a -> b -> c) -> TracedT m a -> TracedT m b -> TracedT m c # (>) :: TracedT m a -> TracedT m b -> TracedT m b # (<*) :: TracedT m a -> TracedT m b -> TracedT m a #
Monad m => Functor (TracedT m) Source #
Instance details Defined in Control.Monad.Bayes.Traced.Static Methods fmap :: (a -> b) -> TracedT m a -> TracedT m b # (<$) :: a -> TracedT m b -> TracedT m a #
Monad m => Monad (TracedT m) Source #
Instance details Defined in Control.Monad.Bayes.Traced.Static Methods (>>=) :: TracedT m a -> (a -> TracedT m b) -> TracedT m b # (>>) :: TracedT m a -> TracedT m b -> TracedT m b # return :: a -> TracedT m a #
MonadDistribution m => MonadDistribution (TracedT m) Source #
Instance details Defined in Control.Monad.Bayes.Traced.Static Methods random :: TracedT m Double Source # uniform :: Double -> Double -> TracedT m Double Source # normal :: Double -> Double -> TracedT m Double Source # gamma :: Double -> Double -> TracedT m Double Source # beta :: Double -> Double -> TracedT m Double Source # bernoulli :: Double -> TracedT m Bool Source # categorical :: Vector v Double => v Double -> TracedT m Int Source # logCategorical :: (Vector v (Log Double), Vector v Double) => v (Log Double) -> TracedT m Int Source # uniformD :: [a] -> TracedT m a Source # geometric :: Double -> TracedT m Int Source # poisson :: Double -> TracedT m Int Source # dirichlet :: Vector v Double => v Double -> TracedT m (v Double) Source #
MonadFactor m => MonadFactor (TracedT m) Source #
Instance details Defined in Control.Monad.Bayes.Traced.Static Methods score :: Log Double -> TracedT m () Source #
MonadMeasure m => MonadMeasure (TracedT m) Source #
Instance details Defined in Control.Monad.Bayes.Traced.Static

hoist :: (forall x. m x -> m x) -> TracedT m a -> TracedT m a Source #

marginal :: Monad m => TracedT m a -> m a Source #

Discard the trace and supporting infrastructure.

mhStep :: MonadDistribution m => TracedT m a -> TracedT m a Source #

A single step of the Trace Metropolis-Hastings algorithm.

mh :: MonadDistribution m => Int -> TracedT m a -> m [a] Source #

Full run of the Trace Metropolis-Hastings algorithm with a specified number of steps. Newest samples are at the head of the list.

For example:

I have forgotten what day it is.
There are ten buses per hour in the week and three buses per hour at the weekend.
I observe four buses in a given hour.
What is the probability that it is the weekend?

>>> :{
 let
   bus = do x <- bernoulli (2/7)
            let rate = if x then 3 else 10
            factor $ poissonPdf rate 4
            return x
   mhRunBusSingleObs = do
     let nSamples = 2
     sampleIOfixed $ unweighted $ mh nSamples bus
 in mhRunBusSingleObs
:}
[True,True,True]

Of course, it will need to be run more than twice to get a reasonable estimate.