I have two matrices, user-rating matrix and item metadata matrix, the item metadata matrix was clustered while the user-rating matrix was not clustered. is it possible to integrate the two together for collaborative filtering (CF) recommendations? If so, how?
hi,
Yes it is possible to combine user-rating matrix with item metadata matrix. But then it used to call as content based recommendation not collaborative filtering based recommendations.
Some of the techniques that can do this are Joint Martix Factorization [1] and User feature Hybrid Model [2]
[1] Kushwaha, N., Mehrotra, S., Kalia, R. et al. Arab J Sci Eng (2016) 41: 5077. https://doi.org/10.1007/s13369-016-2209-0
[2] Symeonidis, Panagiotis, Alexandros Nanopoulos, and Yannis Manolopoulos. "Feature-weighted user model for recommender systems." International Conference on User Modeling. Springer, Berlin, Heidelberg, 2007.
- Badges
- Science topic
- Similar topics
- Computer Science
- Data Mining