"Lemmatized" column could be transformed to word-ordered VW

#5
by Alvant - opened
TopicNet org

Could be useful for intra-text coherence computation, for example.

def f(x):
    words = x.strip('[]').split(', ')
    words = [w.strip("'") for w in words]

    return ' '.join(words)


dataset._data['vw_text'] = (
    dataset._data['id']
    + ' |@lemmatized '
    + dataset._data['lemmatized'].apply(f)
)

Sign up or log in to comment