Colada.WordClass

Word Class induction with LDA

This module provides function which implement word class induction using the generic algorithm implemented in Colada.LDA.

You can access and set options in the Options record using lenses. Example:

  import Data.Label
  let options =   set passes 5 
                . set beta 0.01 
                . set topicNum 100 
                $ defaultOptions
  in run options sentences

Running the sampler

learn :: Options -> [Sentence] -> (WordClass, [Vector (Vector Double)])

learn options xs runs the LDA Gibbs sampler for word classes with options on sentences xs, and returns the resulting model together progressive class the assignments

defaultOptions :: Options

Extracting information

summary :: WordClass -> Text

summary m returns a textual summary of word classes found in model m

summarize :: Bool -> WordClass -> Text

wordTypeClasses :: WordClass -> Map Text (IntMap Double)

wordTypeClasses m returns a Map from word types to unnormalized distributions over word classes

Class and word prediction

label :: Bool -> WordClass -> Sentence -> Vector (Vector Double)

label m s returns for each word in sentences s, unnormalized probabilities of word classes.

predict :: WordClass -> Sentence -> [Vector (Double, Text)]

predict m s returns for each word in sentence s, unnormalized probabilities of words given predicted word class.

Data types and associated lenses

data WordClass

Container for the Word Class model

Instances

Generic WordClass

Serialize WordClass

ldaModel :: forall arr. Arrow arr => Lens arr WordClass Finalized

LDA model

wordTypeTable :: forall arr. Arrow arr => Lens arr WordClass (AtomTable (Vector Char))

Word type string to atom and vice versa conversion tables

featureTable :: forall arr. Arrow arr => Lens arr WordClass (AtomTable (Vector Char))

Feature string to atom and vice versa conversion tables

options :: forall arr. Arrow arr => Lens arr WordClass Options

Options for Gibbs sampling

data Finalized

Instances

Generic Finalized

Serialize Finalized

docTopics :: Finalized -> Table2D

Document topic counts

wordTopics :: Finalized -> Table2D

Word topic counts

topics :: Finalized -> Table1D

Topics counts

topicDocs :: Finalized -> Table2D

Inverse document-topic counts

topicWords :: Finalized -> Table2D

Inverse word-topic counts

data Options

Instances

Eq Options

Data Options

Show Options

Typeable Options

Generic Options

Serialize Options

featIds :: forall arr. Arrow arr => Lens arr Options [Int]

Feature ids

topicNum :: forall arr. Arrow arr => Lens arr Options Int

Number of topics K

alphasum :: forall arr. Arrow arr => Lens arr Options Double

Dirichlet parameter alpha*K which controls topic sparseness

beta :: forall arr. Arrow arr => Lens arr Options Double

Dirichlet parameter beta which controls word sparseness

passes :: forall arr. Arrow arr => Lens arr Options Int

Number of sampling passes per batch

repeats :: forall arr. Arrow arr => Lens arr Options Int

Number of repeats per sentences

batchSize :: forall arr. Arrow arr => Lens arr Options Int

Number of sentences per batch

seed :: forall arr. Arrow arr => Lens arr Options Word32

Random seed for the sampler

topn :: forall arr. Arrow arr => Lens arr Options Int

Number of most probable words to return

initSize :: forall arr. Arrow arr => Lens arr Options Int

initPasses :: forall arr. Arrow arr => Lens arr Options Int

exponent :: forall arr. Arrow arr => Lens arr Options (Maybe Double)

progressive :: forall arr. Arrow arr => Lens arr Options Bool

lambda :: forall arr. Arrow arr => Lens arr Options Double