MSETarget

Bases: Metric

Source code in metrics_toolbox/metrics/regression/mse_target.py 
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
class MSETarget(Metric):
    _name = MetricNameEnum.MSE
    _scope = MetricScopeEnum.TARGET
    _type = MetricTypeEnum.SCORES

    def __init__(self, target_name: str, opt_metadata_series_length: int = 1000):
        """Initialize the Mean Squared Error metric for a specific class.

        Parameters
        ----------
        target_name : str
            The class/column for which to compute the Mean Squared Error.
        opt_metadata_series_length : int, optional
            The length to which the original series values and predicted series values
            will be down sampled in the metadata of the MetricResult.
        """
        self.target_name = target_name
        self.opt_metadata_series_length = opt_metadata_series_length

    @property
    def id(self) -> str:
        """Get the unique identifier for the metric.

        Returns
        -------
        str
            The unique identifier.
        """
        return f"{self.name.value}_{self.target_name}"

    def compute(
        self, y_true: np.ndarray, y_pred: np.ndarray, column_names: list[str]
    ) -> MetricResult:
        """Compute the Mean Squared Error for a specific column in a multi domain
        setting.

        Parameters
        ----------
        y_true : np.ndarray
            True series values for the specified target column.
        y_pred : np.ndarray
            Predicted series values for the specified target column.
        column_names : list[str], optional
            List of column names from model.classes_.

        Returns
        -------
        MetricResult
            The computed Mean Squared Error metric result for the specified target, including
            false down sampled original series values and predicted series values in metadata.
        """

        class_index = column_names.index(self.target_name)

        # Compute the Mean Squared Error for the specified target column
        mse_array = (y_true[:, class_index] - y_pred[:, class_index]) ** 2
        value = mse_array.mean()

        # Down sample the original series values and predicted series values for metadata
        if len(y_true) > self.opt_metadata_series_length:
            indices = np.linspace(
                0, len(y_true) - 1, self.opt_metadata_series_length, dtype=int
            )
            y_true_sampled = y_true[indices, class_index]
            y_pred_sampled = y_pred[indices, class_index]
            mse_array_sampled = mse_array[indices]
        else:
            y_true_sampled = y_true[:, class_index]
            y_pred_sampled = y_pred[:, class_index]
            mse_array_sampled = mse_array

        return MetricResult(
            name=self.name,
            scope=self.scope,
            type=self.type,
            value=value,
            metadata={
                "target_name": self.target_name,
                "y_true": y_true_sampled.tolist(),
                "y_pred": y_pred_sampled.tolist(),
                "error": mse_array_sampled.tolist(),
                "indices": (
                    indices.tolist()
                    if len(y_true) > self.opt_metadata_series_length
                    else None
                ),
            },
        )

id property

Get the unique identifier for the metric.

Returns:
  • str –

    The unique identifier.

__init__(target_name, opt_metadata_series_length=1000)

Initialize the Mean Squared Error metric for a specific class.

Parameters:
  • target_name (str) –

    The class/column for which to compute the Mean Squared Error.

  • opt_metadata_series_length (int, default: 1000 ) –

    The length to which the original series values and predicted series values will be down sampled in the metadata of the MetricResult.
Source code in metrics_toolbox/metrics/regression/mse_target.py 
17
18
19
20
21
22
23
24
25
26
27
28
29
def __init__(self, target_name: str, opt_metadata_series_length: int = 1000):
    """Initialize the Mean Squared Error metric for a specific class.

    Parameters
    ----------
    target_name : str
        The class/column for which to compute the Mean Squared Error.
    opt_metadata_series_length : int, optional
        The length to which the original series values and predicted series values
        will be down sampled in the metadata of the MetricResult.
    """
    self.target_name = target_name
    self.opt_metadata_series_length = opt_metadata_series_length

compute(y_true, y_pred, column_names)

Compute the Mean Squared Error for a specific column in a multi domain setting.

Parameters:
  • y_true (ndarray) –

    True series values for the specified target column.

  • y_pred (ndarray) –

    Predicted series values for the specified target column.

  • column_names (list[str]) –

    List of column names from model.classes_.
Returns:
  • MetricResult –

    The computed Mean Squared Error metric result for the specified target, including false down sampled original series values and predicted series values in metadata.
Source code in metrics_toolbox/metrics/regression/mse_target.py 
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
def compute(
    self, y_true: np.ndarray, y_pred: np.ndarray, column_names: list[str]
) -> MetricResult:
    """Compute the Mean Squared Error for a specific column in a multi domain
    setting.

    Parameters
    ----------
    y_true : np.ndarray
        True series values for the specified target column.
    y_pred : np.ndarray
        Predicted series values for the specified target column.
    column_names : list[str], optional
        List of column names from model.classes_.

    Returns
    -------
    MetricResult
        The computed Mean Squared Error metric result for the specified target, including
        false down sampled original series values and predicted series values in metadata.
    """

    class_index = column_names.index(self.target_name)

    # Compute the Mean Squared Error for the specified target column
    mse_array = (y_true[:, class_index] - y_pred[:, class_index]) ** 2
    value = mse_array.mean()

    # Down sample the original series values and predicted series values for metadata
    if len(y_true) > self.opt_metadata_series_length:
        indices = np.linspace(
            0, len(y_true) - 1, self.opt_metadata_series_length, dtype=int
        )
        y_true_sampled = y_true[indices, class_index]
        y_pred_sampled = y_pred[indices, class_index]
        mse_array_sampled = mse_array[indices]
    else:
        y_true_sampled = y_true[:, class_index]
        y_pred_sampled = y_pred[:, class_index]
        mse_array_sampled = mse_array

    return MetricResult(
        name=self.name,
        scope=self.scope,
        type=self.type,
        value=value,
        metadata={
            "target_name": self.target_name,
            "y_true": y_true_sampled.tolist(),
            "y_pred": y_pred_sampled.tolist(),
            "error": mse_array_sampled.tolist(),
            "indices": (
                indices.tolist()
                if len(y_true) > self.opt_metadata_series_length
                else None
            ),
        },
    )