A data scientist is developing a machine learning model to predict house prices based on features like size, location, and number of bedrooms. After training the model, they notice that the predictions are consistently higher than the actual prices. They decide to use a loss function to evaluate their model's performance. Which loss function would be most appropriate for penalizing these discrepancies effectively?

The most common choice for this situation is the Mean Squared Error (MSE) loss. MSE takes the difference between each predicted price and the true price, squares that difference, and averages over all houses. Because the error is squared, larger mistakes—such as consistently over‑estimating—receive a much heavier penalty, pushing the model to adjust downward. For example, if a house actually sells for $200k but the model predicts $250k, the error is $50k; squaring it gives $2. 5 billion, a large penalty that forces the algorithm to reduce such over‑predictions.