Predicting Mental Health Problems with Automatic Identification of Metaphors

Algorithm 2

Metaphor identification algorithm.

	Input: sentence; A dictionary that returns the corresponding word vector of key, Word2vec; A dictionary that returns the related word set of key, WordNet.
	Output: A list of the corresponding metaphority labels of words in sentence, labels.
(1)	function TokenMetaphorIdentify (detected word; sentence)
(2)	context = [];  = 0
(3)	= WordNet [detected word]
(4)	for each word  sentence and word ≠ detected_word do
(5)	context = context  word
(6)	end for
(7)	= average (Word2vec [context])
(8)	for each k  do
(9)	if cosine(Word2vec[k],) ˃ max_cosine then
(10)	max_cosine = cosine(Word2vec[k],)
(11)	= k
(12)	end if
(13)	end for
(14)	metaphor_value = cosine (Word2vec [detected_word], Word2vec [])
(15)	if metaphor value < threshold then
(16)	return True
(17)	else
(18)	return False
(19)	end if
(20)	end function
(21)
(22)	function MetaphorIdentify (sentence)
(23)	labels = []
(24)	for each word  sentence do
(25)	label = TokenMetaphorIdentify (word; sentence)
(26)	labels = labels  label
(27)	end for
(28)	return labels
(29)	end function