Fork me on GitHub
#data-science
<
2018-01-04
>
elise_huard17:01:54

have people been doing much text mining in clojure? we're looking around for good libs

elise_huard17:01:39

we've developed and are using this https://github.com/MastodonC/kixi.mallet mostly for topic analysis, but keen to know what other people are using

elise_huard17:01:51

(thin wrapper on top of java Mallet)

aaelony20:01:48

I think the cool kids use TensorFlow for text these days, no? That said, perhaps searching for particular methods helps. e.g. a quick search for word2vec led me to https://github.com/Bridgei2i/clojure-word2vec which appears to work on the limited dataset it includes. You can also experiment with deeplearning4j (e.g. https://github.com/deeplearning4j/dl4j-examples/blob/master/dl4j-examples/src/main/java/org/deeplearning4j/examples/nlp/paragraphvectors/ParagraphVectorsClassifierExample.java)

aaelony20:01:46

it would be nice to have seq2seq options in clojure (somehow)...