A Python library for efficient text classification and word representation
fastText
An extension of word2Vec library.
Overcome drawbacks like-
No Sentence representation.
Not exploiting morphology.
Simple to use and much faster.
fastText
Can learn either unsupervised, training on unstructured data or in supervised manner, from series of labelled docs.
Incorporates two research methodologies-
Enriching word vectors with subword info (Unsupervised) .
Bag of tricks for efficient text classification (Supervised) .
Performance is in par with "Deep Learning" algorithms but training time is much lower than them.
fastText
cbow, skip-gram and bow text classification are instances of this model.
cbow - Given many words predict a word.
skip-gram -Given a word predict a word.
bow text classification- Given many words predict a label.
fastText Applications
To classify the text into various categories. e.g. Art Opiate is an online magazine that is published on a quarterly basis and features young artists with their creations. The category it belongs to can be art, magazine etc.
Word representations with character-level features.
fastText Applications
To classify the text into various categories. e.g. Art Opiate is an online magazine that is published on a quarterly basis and features young artists with their creations. The category it belongs to can be art, magazine etc.
Word representations with character-level features.