I am building a recommender system for automatic recommendation of e-learning resources to learners as part of my PhD research. Are there existing algorithms and code for collaborative filtering and association rule mining out there that I can use with python or java?
Weka has an implementation in java for this http://www.cs.waikato.ac.nz/ml/weka/
Some of the state-of-the-art recommender systems are also implemented in Mahout, which you can get for either Hadoop or Spark. You can use user-based recommendation, item-based recommendation and SVD techniques on scores and unary data. The choice would depend on your type of data and feedback from students.
Dear John.
As already mentioned, check out Mahout. Getting started with Mahout can have quite a learning curve. Myself I find LibRec easier to get started with and contains many different algorithsm, though it's a bit academic. MyMediaLite and EasyRec are also good for simple things that do not require Mahout.
For a Java based association rule mining library alternative to Weka's, check out SPMF.
All the best with your research!
http://librec.net/
https://github.com/jcnewell/MyMediaLiteJava/
http://easyrec.org/
http://www.philippe-fournier-viger.com/spmf/
Thanks ladies and gentlemen, you suggestions are very helpful and has given me further insight and more than enough possibilities of solutions.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python