machine learning - mahout for content based recomendation -
i have list user data : user name, age, sex , address, location etc ,
a set of product data : product name, cost , description etc
now build recommendation engine able :
1 figure out similar products
eg :
name : category : cost : ingredients x : x1 : 15 : xx1, xx2, xx3 y : y1 : 14 : yy1, yy2, yy3 z : x1 : 12 : xx1, xy1
here x , z similar.
2 recommend relevant products product list user
how kind or recommendation engine can implement mahout ? available methods ? there useful tutorial/link available ? please help
in mahout v1 here https://github.com/apache/mahout can use "spark-rowsimilarity" create indicators each type of metadata, categroy, cost, , ingredients. give 3 matrices containing similar items each item based on particular metadata. give "more this" type of recommendation. can try combining metadata 1 input matrix , see if gives better results.
to personalize record items user has expressed preference for. index indicator matrices in solr, 1 indicator per solr "field" attached item id (name?). query user's history against each field. can boost fields increase weight in recommendations.
this described on mahout site: http://mahout.apache.org/users/recommender/intro-cooccurrence-spark.html , slides here: http://occamsmachete.com/ml/2014/10/07/creating-a-unified-recommender-with-mahout-and-a-search-engine/
Comments
Post a Comment