1


I am new to apache mahout but i read one article which said Apache Mahout 1.0 gives content based recommendion (http://mahout.apache.org/users/algorithms/intro-cooccurrence-spark.html) but now it turns out that it does not give content-based recommendation rather it gives recommendation based on different user actions on website.

Amazon ,Netflix might have been using content-based recommender and probabily they might have implemented them from scratch but now my question is:

Is there any Machine Learning library which gives us content-based recommendation or do i have to implement it by myself ?
Here by content based recommendation i mean there is feature vector for item and we behaviour vector for each user and hence by multiplying them we get recommendation for particular user.

Please recommend something to me,
Thanks in advance.

KlwntSingh
  • 1,084
  • 8
  • 26
  • That page shows one method of making "content-based" recommendations. Look at "spark-rowsimilairty". Please edit your question with more information about what you mean by "content". Amazon and Netflix use collaborative filtering for recommendations--not content-based. Therefore they use user "preferences". Amazon uses user purchases as implicit preferences. These methods are covered in the Mahout page. – pferrel Apr 16 '15 at 17:27

0 Answers0