Tutorial on application-oriented evaluation of recommendation systems Academic Article uri icon

abstract

  • Abstract Recommendation systems are now widely used in many commercial applications. This tutorial focuses on the evaluation of such systems, from an application-oriented view. The tutorial recommends best practices, suggests a protocol for the evaluation process, and reviews a set of metrics that can be evaluated. The practices in this paper are motivated by similar procedures in nearby areas, such as machine learning, and information retrieval.

publication date

  • January 1, 2013