You are training a machine learning model to classify customer reviews as positive or negative. After defining the model, you want to prevent overfitting by carefully monitoring the training process. Which of the following techniques is the most effective in preventing overfitting and ensuring generalization of the model to new data?