• tfidf_rank

  • features: TF, IDF (==1)
  • ranking: TF * IDF

• text_rank

  • features: pos (for filtering), word neighbors
  • ranking: TextRank, which like PageRank, while building a relation matrix according to the words' positions.
  • reference: http://www.cse.unt.edu/~rada/papers/mihalcea.emnlp04.pdf

• glm_rank (TODO)

  • features: word frequence (TF), word importance (IDF, Part-Of-Speech, Entity type), word position (such as whether both in title and body)
  • ranking: train and predict by regression model.

• semantic_rank (TODO)

  • features: such as TF, IDF, POS, entity type ...
  • ranking: regression model plus SemanticRank, which like Pagerank, while building a relation matrix according to semantic similarity.
  • re-ranking: document category based adjusting, task dependent word adjusting.

• topic_rank (TODO)

  • ranking: topic model, such as LDA.

##Requirements:

• jieba: chinese word segmenter, https://github.com/fxsjy/jieba

• word2vec: for building word similarities, https://code.google.com/p/word2vec/