Recommendation Engines
< prev next >

Behavioral recommendation engines are conceptually easy to implement: if user X liked items 5,6, and 7, user Z which liked items 5 and 7, will probably like item 6.

Now, if you have millions of items and millions of users, a complex software engineering problem.

Base 10 Labs’ recommendation engine makes use of an efficient sparse matrix library capable of handling millions of items, in many categories, and with fallback rules to return recommendations even for items or users which have very few data points to compare.