java - content based recommender in hadoop -
i thinking implementing content based recommendation project idea in hadoop.
just wanted suggestions if weka can used hadoop or if advisable use mahout cbf based implementation
please advise
thanks
what content? data surrounds or makes items?
there several ways use recs , can blend in collaborative filtering data if appropriate. take @ rowsimilarity in mahout. calculates similarity of item others. if input content in way can digest you'll have cbr.
for instance input recommending similar blogs posts be: postid, tokens-from-the-text-of-the-post
then each postid you'll list other postids use similar words. can same tags, categories, or other metadata too.
Comments
Post a Comment