no code implementations • 2 Jun 2014 • Saima Aman, Yogesh Simmhan, Viktor K. Prasanna
The performance of prediction models is often based on "abstract metrics" that estimate the model's ability to limit residual errors between the observed and predicted values.