Evaluation

Evaluation is a process that assesses some property of an artifact. In machine learning, two types of∖break artifacts are most commonly evaluated, models and {algorithms}. Model evaluation often focuses on the predictive efficacy of the model, but may also assess factors such as its complexity, the ease with which it can be understood, or the computational requirements for its application. Algorithm evaluation often focuses on evaluation of the models an algorithm produces, but may also appraise its computational efficiency.