src/Gargantext/Text/Metrics/TFICF.hs

   1 {-|
   2 Module      : Gargantext.Text.Metrics.TFICF
   3 Description : TFICF Ngrams tools
   4 Copyright   : (c) CNRS, 2017
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 Definition of TFICF : Term Frequency - Inverse of Context Frequency
  11
  12 -}
  13
  14 {-# LANGUAGE NoImplicitPrelude #-}
  15 {-# LANGUAGE OverloadedStrings #-}
  16
  17 module Gargantext.Text.Metrics.TFICF where
  18
  19 --import Data.Text (Text)
  20 import Gargantext.Prelude
  21 import Gargantext.Database.Schema.Ngrams (NgramsId, NgramsTerms)
  22
  23 data TficfContext n m = TficfLanguage n m | TficfCorpus n m | TficfDocument n m
  24   deriving (Show)
  25
  26 data Tficf = Tficf { tficf_ngramsId :: NgramsId
  27                    , tficf_ngramsTerms :: NgramsTerms
  28                    , tficf_score       :: Double
  29 } deriving (Show)
  30
  31
  32 type SupraContext = TficfContext
  33 type InfraContext = TficfContext
  34
  35 -- | TFICF is a generalization of TFIDF
  36 -- https://en.wikipedia.org/wiki/Tf%E2%80%93idf
  37 tficf :: InfraContext Double Double -> SupraContext Double Double -> Double
  38 tficf (TficfCorpus c c')  (TficfLanguage l l') = tficf' c c' l l'
  39 tficf (TficfDocument d d')(TficfCorpus   c c') = tficf' d d' c c'
  40 tficf _ _ = panic "Not in definition"
  41
  42 tficf' :: Double -> Double -> Double -> Double -> Double
  43 tficf' c c' l l'
  44     | c <= c' && l < l' = (l/l') / log (c/c')
  45     | otherwise        = panic "Frequency impossible"
  46
  47
  48 tficf_example :: [(Double,Double,Double,Double)]
  49 tficf_example = undefined