After training a regression model, a specialist wants to determine whether predictions tend to be systematically above or below the true values. Which diagnostic method will show whether the model is overestimating or underestimating the target?

Residual plots are the correct choice because they graphically display the difference between predicted and actual values (residuals) against the predicted values or independent variables. This visualization helps identify patterns, such as a systematic upward or downward trend, indicating whether the model consistently overestimates or underestimates the target. RMSE provides an overall measure of prediction error but doesn't show the direction of errors. Area under the curve (AUC) is used for classification problems, not regression. A confusion matrix is also for classification tasks, summarizing true positives, true negatives, false positives, and false negatives.