smape

smape(y_true: Union[float, Sequence[float], Sequence[Sequence[float]]], y_pred: Union[float, Sequence[float], Sequence[Sequence[float]]], eps: float = 1e-15, multioutput: str = 'joint') Union[float, Sequence[float], Sequence[Sequence[float]]][source]

Symmetric mean absolute percentage error.

Wikipedia entry on the Symmetric mean absolute percentage error

\[SMAPE = \dfrac{100}{n}\sum_{t=1}^{n}\dfrac{|ytrue_{t}-ypred_{t}|}{(|ypred_{t}|+|ytrue_{t}|) / 2}\]
Parameters
  • y_true (Union[float, Sequence[float], Sequence[Sequence[float]]]) –

    array-like of shape (n_samples,) or (n_samples, n_outputs)

    Ground truth (correct) target values.

  • y_pred (Union[float, Sequence[float], Sequence[Sequence[float]]]) –

    array-like of shape (n_samples,) or (n_samples, n_outputs)

    Estimated target values.

  • eps (float=1e-15) – SMAPE is undefined for y_true[i] + y_pred[i] == 0 for any i, so all zeros y_true[i] + y_pred[i] are clipped to max(eps, abs(y_true) + abs(y_pred)).

  • multioutput (str) – Defines aggregating of multiple output values (see FunctionalMetricMultioutput).

Returns

A non-negative floating point value (the best value is 0.0), or an array of floating point values, one for each individual target.

Return type

Union[float, Sequence[float], Sequence[Sequence[float]]]