|
What are some standard evaluation methods for a classification system, such as hand written recognition using MNIST database? Currently I can only think of training error rate ( obtained through cross validation) and test error rate. Any other important metric? |
