Suppose we have a dataset with 500 spam emails and 500 non-spam emails. When we apply our learned model to this dataset, suppose our model correctly predicts 300 of the 500 spam emails as spam, and incorrectly predicts 100 of the 500 non-spam emails as spam, what is the precision and recall of our spam detector. Suppose we change our predictor to label all 1000 emails as spam, which value would be higher precision or recall