3 ms·
Interesting. I had heard specifically that word vectors weren't a game-changer for document classification, because the averaging method didn't work well. But t
by ppod 9y ago
Interesting. I had heard specifically that word vectors weren't a game-changer for document classification, because the averaging method didn't work well. But the difference between word2vec and FastText must be important. I think your baseline of Naive Bayes + TF-IDF would do a lot better if it was SVM and TF-IDF, but I didn't expect the big jump between word2vec and FastText.
- wenc 9y ago> I had heard specifically that word vectors weren't a game-changer for document classification, because the averaging method didn't work well. As with anything, your mileage may vary. One aspect of FastText that definitely helped in my case was n-gram support (both word and character, tunable via command-line arguments). In my corpus, I have short fragments of sentences containing misspelled words, incorrect grammar etc. plus my test set has out-of-vocabulary words. n-grams are more robust to these than Word2Vec which uses a static vocabulary.
- swsieber 9y agoI thanked your top level comment, but this whole thread is great. I fond my self worth data that sounds like yours. I'm excited to try fast text.
- sp332 9y agoI think it's important to use the right "average" operation. word2vec puts words on the surface of a hypersphere, so "adding" vectors means turning around the sphere.