Your deep learning model for a classification task is showing low precision and recall scores. Which of the following strategies could potentially enhance these metrics?