Mean Absolute Error

A metric for regression tasks that measures the average absolute difference between predicted and actual values. MAE is easy to interpret, telling you in plain units how far off predictions are on average, and is less sensitive to large outliers than mean squared error.