NLP.Concraft.DAG.Disamb

Types

data Disamb t

A disambiguation model.

Constructors

Disamb

Fields

• tiers :: [Tier]
• schemaConf :: SchemaConf
• crf :: CRF Ob Atom
• simplify :: t -> Tag

  A map which simplifies the tags of generic type t to simplified positional tags. The motivation behind this is that tags can have a richer structure.

  NOTE: it can happen in real situations that a tag is encountered which is not known by the model. It would be nice to be able to treat it as the closest tag that can be handled. Then, one have to define the notion of the similarilty between tags, though... But probably it should be done at a different level (where more information about the structure of t is known)

putDisamb :: Disamb t -> Put

Store the entire disambiguation model apart from the simplification function.

getDisamb :: (t -> Tag) -> Get (Disamb t)

Get the disambiguation model, provided the simplification function. getDisamb :: (M.Map t T.Tag) -> Get (Disamb t)

Tiers

data Tier

A tier description.

Constructors

Tier

Fields

• withPos :: Bool

  Does it include the part of speech?

• withEos :: Bool

  End-of-sentence marker.

• withAtts :: Set Attr

  Tier grammatical attributes.

Instances

Binary Tier

Defined in NLP.Concraft.Disamb.Positional

Methods

put :: Tier -> Put

get :: Get Tier

putList :: [Tier] -> Put

data Atom

An atomic part of morphosyntactic tag with optional POS.

Constructors

Atom

Fields

• pos :: Maybe POS
• atts :: Map Attr Text
• eos :: Maybe Bool

  NOTE: could be simplified to Bool, but this way it's more readable

Instances

Eq Atom

Defined in NLP.Concraft.Disamb.Positional

Methods

(==) :: Atom -> Atom -> Bool

(/=) :: Atom -> Atom -> Bool

Ord Atom

Defined in NLP.Concraft.Disamb.Positional

Methods

compare :: Atom -> Atom -> Ordering

(<) :: Atom -> Atom -> Bool

(<=) :: Atom -> Atom -> Bool

(>) :: Atom -> Atom -> Bool

(>=) :: Atom -> Atom -> Bool

max :: Atom -> Atom -> Atom

min :: Atom -> Atom -> Atom

Show Atom

Defined in NLP.Concraft.Disamb.Positional

Methods

showsPrec :: Int -> Atom -> ShowS

show :: Atom -> String

showList :: [Atom] -> ShowS

Binary Atom

Defined in NLP.Concraft.Disamb.Positional

Methods

put :: Atom -> Put

get :: Get Atom

putList :: [Atom] -> Put

Probs in general

data ProbType

Type of resulting probabilities.

Constructors

Marginals

Marginal probabilities

MaxProbs

TODO

probsSent :: (Word w, Ord t) => ProbType -> Disamb t -> Sent w t -> Sent w t

Determine the marginal probabilities of to individual labels in the sentence. marginalsSent :: (X.Word w, Ord t) => Disamb t -> X.Sent w t -> DAG () (X.WMap [P.Atom])

probs :: (Word w, Ord t) => ProbType -> Disamb t -> Sent w t -> DAG () (WMap t)

Determine the marginal probabilities of to individual labels in the sentence.

Training

data TrainConf t

Training configuration.

Constructors

TrainConf

Fields

• tiersT :: [Tier]
• schemaConfT :: SchemaConf
• sgdArgsT :: SgdArgs
• onDiskT :: Bool
• simplifyLabel :: t -> Tag

  Label simplification function

train

Arguments

:: (Word w, Ord t)

=> TrainConf t

Training configuration

-> IO [Sent w t]

Training data

-> IO [Sent w t]

Evaluation data

-> IO (Disamb t)

Train disambiguation module.

Pruning

prune :: Double -> Disamb t -> Disamb t

Prune disamb model: discard model features with absolute values (in log-domain) lower than the given threshold.

Internal

schematize :: Schema w [t] a -> Sent w [t] -> Sent Ob t

Schematize the input sentence according to schema rules.