pythongraphlab

Collaborative filtering in Python


I work now with the Graphlab. After installing of this program, I could run the algorithms of Collaborative filtering. Now I try to work with Graphlab in the Python. I have found already this brilliant toolkits. My question is, does anybody know about Collaborative Filtering in the python implementation? I couldn't find this tool & it seems me it doesn't implemented for Python...

Thank you in advance.


Solution

  • Check out the recommender package in GraphLab Create. It lets you create a collaborative filtering model in just a few lines.

    import graphlab
    sf = graphlab.SFrame.read_csv('my_data.csv')
    m = graphlab.recommender.create(data)
    recs = m.recommend()
    

    You will likely be most interested in the item similarity models, but you should also check out the other options for the method argument, such as matrix_factorization.