Using the already defined RBF SVC model `m`, run two grid searches on the parameters C and gamma, for values [0.05, 0.5, 1, 5, 20, 100, 500]. The first grid search should find the model that best optimizes for recall. The second grid search should find the model that best optimizes for precision. How much better is the best recall (found by first grid search) than the best precision you found (result of second grid search)? (Compute the difference between these two as: recall - precision, to 3 decimal places)

(Training data are in X_train, y_train, and use y_test and X_test to compute precision and recall.)

1

2