Skip to content

Latest commit

 

History

History
24 lines (15 loc) · 1.37 KB

README.md

File metadata and controls

24 lines (15 loc) · 1.37 KB

entity2rec

Implementation of the entity recommendation algorithm described in entity2rec: Learning User-Item Relatedness from Knowledge Graphs for Top-N Item Recommendation. Compute user and item embeddings from a Knowledge Graph encompassing both user feedback information and item information. It is based on property-specific entity embeddings, which are obtained via entity2vec (https://github.com/MultimediaSemantics/entity2vec). Slides can be found on Slideshare. The main difference between the current implementation and what is reported in the paper is the evaluation protocol, which now ranks all the items for each user.

For a usage example, see the Wiki section.

Requirements

  • Python 2.7 or above
  • numpy
  • gensim
  • networkx 1.x
  • pandas
  • SPARQL Wrapper

If you are using pip:

    pip install gensim networkx pandas SPARQLWrapper

Our Publications