Metrics

The gtime.metrics module contains a collection of different metrics.

gtime.metrics.gmae(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray]) → float

Compute the Geometric Mean Absolute Error between two vectors.

Parameters

y_truearray-like, shape (length, 1), required: The first vector.
y_predarray-like, shape (length, 1), required: The second vector.

Returns

gmae_valuefloat: The geometric mean absolute error between the two vectors.

Examples

>>> from gtime.metrics import gmae
>>> y_true = [0, 1, 2, 3, 6, 5]
>>> y_pred = [-1, 4, 5, 10, 4, 1]
>>> gmae(y_true, y_pred)
2.82

gtime.metrics.log_mse(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray]) → float

Compute the Mean Squared Log Error(MSLE) between two vectors. Note: Log_mse accepts only positive numbers as input

Parameters

y_truearray-like, shape (length, 1), required: The first vector.
y_predarray-like, shape (length, 1), required: The second vector.

Returns

log_msefloat: The mean squared log error between the two vectors.

Examples

>>> from gtime.metrics import log_mse
>>> y_true = [0, 1, 2, 3, 4, 5]
>>> y_pred = [1.1, 2.3, 0.4, 3.9, 3.1, 4.6]
>>> log_mse(y_true, y_pred)
0.244

gtime.metrics.mae(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray]) → float

Compute the Mean Absolute Error(also called, Mean Absolute Deviation(MAD) or Mean Ratio) between two vectors.

Parameters

y_truearray-like, shape (length, 1), required: The first vector.
y_predarray-like, shape (length, 1), required: The second vector.

Returns

mae_valuefloat: The mean absolute error between the two vectors.

Examples

>>> from gtime.metrics import mae
>>> y_true = [0, 1, 2, 3, 4, 5]
>>> y_pred = [1.1, 2.3, 0.4, 3.9, 3.1, 4.6]
>>> mae(y_true, y_pred)
1.033

gtime.metrics.mape(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray]) → float

Compute the Mean Absolute Percentage Error(MAPE) between two vectors.

Parameters

y_truearray-like, shape (length, 1), required.: The first vector.
y_predarray-like, shape (length, 1), required.: The second vector.

Returns

mape_valuefloat: The mean absolute percentage error between the two vectors.

Examples

>>> from gtime.metrics import mape
>>> y_true = [1, 1, 2, 3, 4, 5]
>>> y_pred = [1, 2.3, 0.4, 3.9, 3.1, 4.6]
>>> mape(y_true, y_pred)
45.08

gtime.metrics.max_error(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray]) → float

Compute the maximum error between two vectors.

Parameters

y_truearray-like, shape (length, 1), required: The first vector.
y_predarray-like, shape (length, 1), required: The second vector.

Returns

errorfloat: The maximum error between the two vectors.

Examples

>>> from gtime.metrics import max_error
>>> y_true = [0, 1, 2, 3, 4, 5]
>>> y_pred = [1.1, 2.3, 0.4, 3.9, 3.1, 4.6]
>>> max_error(y_true, y_pred)
1.6

gtime.metrics.mse(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray]) → float

Compute the Mean Squared Error(MSE) between two vectors.

Parameters

y_truearray-like, shape (length, 1), required.: The first vector.
y_predarray-like, shape (length, 1), required.: The second vector.

Returns

msefloat: The Mean Squared Error between the two vectors.

Examples

>>> from gtime.metrics import mse
>>> y_true = [0, 1, 2, 3, 4, 5]
>>> y_pred = [1.1, 2.3, 0.4, 3.9, 3.1, 4.6]
>>> mse(y_true, y_pred)
1.20

gtime.metrics.non_zero_smape(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray], raise_error: bool = False) → float

Compute the ‘Symmetric Mean Absolute Percentage Error’ (SMAPE) between two vectors without considering the 0 in the true values. Documentation here <https://en.wikipedia.org/wiki/Symmetric_mean_absolute_percentage_error>_.

Parameters

y_truearray-like, shape (length, 1), required: The first vector.
y_predarray-like, shape (length, 1), required: The second vector.

raise_error: bool, optional, default=``False``

Returns

smapefloat: The smape between the two input vectors.

Examples

>>> from gtime.metrics import non_zero_smape
>>> y_true = [0, 1, 2, 3, 4, 5]
>>> y_pred = [1.1, 2.3, 0.4, 3.9, 3.1, 4.6]
>>> non_zero_smape(y_true, y_pred)
0.7864893577539014

gtime.metrics.r_square(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray]) → float

Compute the R squared (Coefficient of Determination) between two vectors.

Parameters

y_truearray-like, shape (length, 1), required: The first vector.
y_predarray-like, shape (length, 1), required: The second vector.

Returns

r_squarefloat: The R squared between the two vectors.

Examples

>>> from gtime.metrics import r_square
>>> y_true = [0, 1, 2, 3, 4, 5]
>>> y_pred = [1.1, 2.3, 0.4, 3.9, 3.1, 4.6]
>>> r_square(y_true, y_pred)
0.586

gtime.metrics.rmse(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray]) → float

Compute the Root Mean Squared Error(RMSE) between two vectors.

Parameters

y_truearray-like, shape (length, 1), required.: The first vector.
y_predarray-like, shape (length, 1), required.: The second vector.

Returns

rmsefloat: The Root Mean Squared Error between the two vectors.

Examples

>>> from gtime.metrics import rmse
>>> y_true = [0, 1, 2, 3, 4, 5]
>>> y_pred = [1.1, 2.3, 0.4, 3.9, 3.1, 4.6]
>>> rmse(y_true, y_pred)
1.098

gtime.metrics.rmsle(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray]) → float

Compute the Root Mean Squared Log Error(RMSLE) between two vectors. Note: RMSLE accepts only positive numbers as input

Parameters

y_truearray-like, shape (length, 1), required.: The first vector.
y_predarray-like, shape (length, 1), required.: The second vector.

Returns

rmsle_valuefloat: The root mean squared log error between the two vectors.

Examples

>>> from gtime.metrics import rmsle
>>> y_true = [0, 1, 2, 3, 4, 5]
>>> y_pred = [1.1, 2.3, 0.4, 3.9, 3.1, 4.6]
>>> rmsle(y_true, y_pred)
0.49

gtime.metrics.smape(y_true: Union[DataFrame, List, ndarray], y_pred: Union[DataFrame, List, ndarray]) → float

Compute the ‘Symmetric Mean Absolute Percentage Error’ (SMAPE) between two vectors. Documentation here <https://en.wikipedia.org/wiki/Symmetric_mean_absolute_percentage_error>_.

Parameters

y_truearray-like, shape (length, 1), required: The first vector.
y_predarray-like, shape (length, 1), required: The second vector.

Returns

smapefloat: The smape between the two input vectors.

Examples

>>> from gtime.metrics import smape
>>> y_true = [0, 1, 2, 3, 4, 5]
>>> y_pred = [1.1, 2.3, 0.4, 3.9, 3.1, 4.6]
>>> smape(y_true, y_pred)
0.7864893577539014