mae

mbgdml.losses.mae(errors)

Mean absolute error (MAE).

\[MAE = \frac{1}{n} \sum_{i=1}^n \mid \hat{y}_{i} - y_{i} \mid,\]

where \(y_{i}\) is the true value, \(\hat{y}_{i}\) is the predicted value, \(\mid \cdots \mid\) is the absolute value, and \(n\) is the number of data points.

Parameters:

errors (numpy.ndarray) – Array of \(\hat{y} - y\) values. This will be flattened beforehand.

Returns:

Mean absolute error.

Return type:

float

See also

mse, rmse, sse