Definition
A regression metric that calculates the average of the absolute differences between the model's predicted values and the actual observed values, treating all errors equally.
In short: A regression metric that calculates the average of the absolute differences between the model's predicted values and the actual observed values, treating all errors equally.