I am trying to annotate the vocabulary in the corpus.
I have trained the word2vec model on the corpus
I have grouped the words which are related based on the score as key as the first word as the key and remaining words as a list of 2-tuple of word and scores with respect to the key
example: 'coffee'---key values are
[('tea', 0.8139282), ('latte', 0.76456803), ('coffe', 0.7607962), ('lattes', 0.756057), ('starbucks', 0.7158153), ('espresso', 0.71386236), ('mocha', 0.69999266), ('coffees', 0.6816252), ('frappucino', 0.67192864), ('cuppa', 0.66720986), ('cappucino', 0.6664002), ('chai', 0.6623157), ('decaf', 0.65980726), ('frappuccino', 0.65150374), ('venti', 0.6486204), ('expresso', 0.6369579), ('macchiato', 0.6280453), ('scone', 0.62476856), ('sippy', 0.6236704), ('cappuccino', 0.61718297), ('iced', 0.6130485), ('hazelnut', 0.6023698), ('mug', 0.6004759), ' ' ' ' '
as i know the coffee is releated to latte ,green_tea ,espresso,starbucks.. from the above data I would like to label each word as below
latte [COHYPO] green_tea [COHYPO] espresso [HYPO] Starbucks [RELATED] tim_horton [RELATED]
[RELATED] -the word is repeated
[MORPHO]-Morphological variant (example :Computer and computers )
[Partof]- indicates that the annotated word is a part of the word of interest
Any suggestion or ideas by which I can approach this problem