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 TFICF is a generalization of [TFIDF](https://en.wikipedia.org/wiki/Tf%E2%80%93idf).
  13
  14 -}
  15
  16 {-# LANGUAGE NoImplicitPrelude #-}
  17 {-# LANGUAGE OverloadedStrings #-}
  18
  19 module Gargantext.Text.Metrics.TFICF ( TFICF
  20                                      , TficfContext(..)
  21                                      , Total(..)
  22                                      , Count(..)
  23                                      , tficf
  24                                      )
  25   where
  26
  27 import Data.Text (Text)
  28 import Gargantext.Prelude
  29
  30 path :: Text
  31 path = "Gargantext.Text.Metrics.TFICF"
  32
  33 type TFICF = Double
  34
  35 data TficfContext n m = TficfInfra n m
  36                       | TficfSupra n m
  37   deriving (Show)
  38
  39 data Total = Total {unTotal :: !Double}
  40 data Count = Count {unCount :: !Double}
  41
  42 tficf :: TficfContext Count Total
  43       -> TficfContext Count Total
  44       -> TFICF
  45 tficf (TficfInfra (Count ic) (Total it) )
  46       (TficfSupra (Count sc) (Total st) )
  47             | it >= ic && st >= sc = (ic/it) / log (sc/st)
  48             | otherwise            = panic $ "[ERR]" <> path <>" Frequency impossible"
  49 tficf _ _ = panic $ "[ERR]" <> path <> "Undefined for these contexts"
  50
  51