August 2018
Intermediate to advanced
522 pages
12h 45m
English
The first approach is based on the Singular Value Decomposition (SVD) of the user-item matrix. This technique allows the transforming of a matrix through a low-rank factorization and can also be used in an incremental way, as described in Incremental Singular Value Decomposition Algorithms for Highly Scalable Recommender Systems, Sarwar B, Karypis G, Konstan J, Riedl J, 2002. In particular, if the user-item matrix has m rows and n columns:

We have assumed that we have real matrices (which is often true in our case), but in general they are complex. U and V are unitary, while Σ is a rectangular diagonal ...
Read now
Unlock full access