fastText Overview!

fastText

A Python library for efficient text classification and word representation

fastText

• An extension of word2Vec library.
• Overcome drawbacks like-

1. No Sentence representation.
2. 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-

1. Enriching word vectors with subword info (Unsupervised) .
2. 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.

​

fastText

• To classify the text into various categories.

​​