Word2vec

gensim

• https://rare-technologies.com/word2vec-tutorial/
• document similarity
  • https://rare-technologies.com/performance-shootout-of-nearest-neighbours-contestants/
  • Using gensim’s memory-friendly streaming API I then converted these plain text tokens to TF-IDF vectors, ran Singular Value Decomposition (SVD) on this TF-IDF matrix to build a latent semantic analysis (LSA) model and finally stored each Wikipedia document as a 500-dimensional LSA vector to disk.

pretrained korean word2vec

• https://github.com/Kyubyong/wordvectors

memo

• https://medium.com/@klintcho/doc2vec-tutorial-using-gensim-ab3ac03d3a1
• wikipedia dump
  • https://dumps.wikimedia.org/backup-index.html

related items

• Singular value decomposition
• FastText
• 나무위키 코퍼스

computational resource

• https://drive.google.com/file/d/0B8XXo8Tve1cxWTVTXzdNTlV4ek0/view
• https://fasttext.cc/docs/en/pretrained-vectors.html#content