(1) Take a list of existing tags and try to identify the type of word. Initial hypothesis that tags would have majority of adjectives, common nouns & Verbs
(2) See distribution of tags into categories
(3) Try following text processing approaches - tf-idf, cosine similarity, cosine similarity with POS tagging and build a supervised classification to categories to see which approach works best to be able to categorise
(1) Take a list of existing tags and try to identify the type of word. Initial hypothesis that tags would have majority of adjectives, common nouns & Verbs (2) See distribution of tags into categories (3) Try following text processing approaches - tf-idf, cosine similarity, cosine similarity with POS tagging and build a supervised classification to categories to see which approach works best to be able to categorise