Large sparse matrices
In commercial implementations of these recommender systems, the utility and similarity matrices would be far too large to be stored as internal arrays. Amazon, for example, has millions of items for sale and hundreds of millions of customers. With m = 100,000,000 and n = 1,000,000, the utility matrix would have m·n = 100,000,000,000,000 slots and the similarity matrix would have n2 = 1,000,000,000,000 slots. Moreover, if the average customer buys 100 items, then only 100n = 100,000,000 of the entries of the utility matrix would be non-zero—that's only 0.0001 percent of the entries, making it a very sparse matrix.
A sparse matrix is a matrix in which nearly all the entries are zero. Even if possible, it is very inefficient ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access