seqlearn icon indicating copy to clipboard operation
seqlearn copied to clipboard

How to use word embedding as a feature ?

Open somnathasati opened this issue 7 years ago • 2 comments

I want to know, if there exist any way to pass word vector extracted from word2vec as a feature ? or, i have to use/change something else in place of existing feature hashing ? Great Thanks !

somnathasati avatar Jun 23 '17 08:06 somnathasati

There is no feature hashing, seqlearn estimators require feature matrix X which can contain arbitrary features, including word vectors.

kmike avatar Jun 23 '17 10:06 kmike

There is no feature hashing, seqlearn estimators require feature matrix X which can contain arbitrary features, including word vectors.

@kmike Hi, does it mean I can directly use feature matrix X, even if it is not sparse, without feature hashing? Thanks a lot!

warshallrho avatar Apr 29 '19 09:04 warshallrho