Accuracy on the test set

$1$ point possiale $($graded$)$

After you have chosen your best method $($perceptron$,$ average perceptron or Pegasos$)$ and parameters, use this classifier to compute testing accuracy on the test set.

We have supplied the feature matrix and labels in $$ and te test labels

Note: In practice the valdation set is used for tuning hyperparameters while a heldout test set is the final benchmark used to compare disparate moldels that have already been tuned. You may notice that your results using a validation set don't always aign with those of the test set, and this is to be expected,

Accuracy on the test set :