Skip to content

Statistics

This section documents the statistics components of the tree module.

statistics

Scikit-learn tree module statistics interoperability for Nextmv.

This module provides functionality to integrate scikit-learn tree-based models with Nextmv statistics tracking.

FUNCTION DESCRIPTION
DecisionTreeRegressorStatistics

Convert a DecisionTreeRegressor model to Nextmv statistics format.

DecisionTreeRegressorStatistics 

DecisionTreeRegressorStatistics(
    model: DecisionTreeRegressor,
    X: Iterable,
    y: Iterable,
    sample_weight: float = None,
    run_duration_start: Optional[float] = None,
) -> Statistics

Create a Nextmv statistics object from a scikit-learn DecisionTreeRegressor model.

You can import the DecisionTreeRegressorStatistics function directly from tree:

from nextmv_sklearn.tree import DecisionTreeRegressorStatistics

Converts a trained scikit-learn DecisionTreeRegressor model into Nextmv statistics format. The statistics include model depth, feature importances, number of leaves, and model score. Additional custom metrics can be added by the user after this function returns. The optional run_duration_start parameter can be used to track the total runtime of the modeling process.

PARAMETER DESCRIPTION

model

The trained scikit-learn DecisionTreeRegressor model.

TYPE: DecisionTreeRegressor

X

The input features used for scoring the model.

TYPE: Iterable

y

The target values used for scoring the model.

TYPE: Iterable

sample_weight

The sample weights used for scoring, by default None.

TYPE: float DEFAULT: None

run_duration_start

The timestamp when the model run started, typically from time.time(), by default None.

TYPE: float DEFAULT: None

RETURNS DESCRIPTION
Statistics

A Nextmv statistics object containing model performance metrics.

Examples:

>>> from sklearn.tree import DecisionTreeRegressor
>>> from nextmv_sklearn.tree import DecisionTreeRegressorStatistics
>>> import time
>>>
>>> # Record start time
>>> start_time = time.time()
>>>
>>> # Train model
>>> model = DecisionTreeRegressor(max_depth=5)
>>> model.fit(X_train, y_train)
>>>
>>> # Create statistics
>>> stats = DecisionTreeRegressorStatistics(
...     model, X_test, y_test, run_duration_start=start_time
... )
>>>
>>> # Add additional metrics
>>> stats.result.custom["my_custom_metric"] = custom_value
Source code in nextmv-scikit-learn/nextmv_sklearn/tree/statistics.py 
 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
100
101
102
103
def DecisionTreeRegressorStatistics(
    model: tree.DecisionTreeRegressor,
    X: Iterable,
    y: Iterable,
    sample_weight: float = None,
    run_duration_start: Optional[float] = None,
) -> nextmv.Statistics:
    """Create a Nextmv statistics object from a scikit-learn DecisionTreeRegressor model.

    You can import the `DecisionTreeRegressorStatistics` function directly from `tree`:

    ```python
    from nextmv_sklearn.tree import DecisionTreeRegressorStatistics
    ```

    Converts a trained scikit-learn DecisionTreeRegressor model into Nextmv statistics
    format. The statistics include model depth, feature importances, number of leaves,
    and model score. Additional custom metrics can be added by the user after this
    function returns. The optional `run_duration_start` parameter can be used to track
    the total runtime of the modeling process.

    Parameters
    ----------
    model : tree.DecisionTreeRegressor
        The trained scikit-learn DecisionTreeRegressor model.
    X : Iterable
        The input features used for scoring the model.
    y : Iterable
        The target values used for scoring the model.
    sample_weight : float, optional
        The sample weights used for scoring, by default None.
    run_duration_start : float, optional
        The timestamp when the model run started, typically from time.time(),
        by default None.

    Returns
    -------
    nextmv.Statistics
        A Nextmv statistics object containing model performance metrics.

    Examples
    --------
    >>> from sklearn.tree import DecisionTreeRegressor
    >>> from nextmv_sklearn.tree import DecisionTreeRegressorStatistics
    >>> import time
    >>>
    >>> # Record start time
    >>> start_time = time.time()
    >>>
    >>> # Train model
    >>> model = DecisionTreeRegressor(max_depth=5)
    >>> model.fit(X_train, y_train)
    >>>
    >>> # Create statistics
    >>> stats = DecisionTreeRegressorStatistics(
    ...     model, X_test, y_test, run_duration_start=start_time
    ... )
    >>>
    >>> # Add additional metrics
    >>> stats.result.custom["my_custom_metric"] = custom_value
    """

    run = nextmv.RunStatistics()
    if run_duration_start is not None:
        run.duration = time.time() - run_duration_start

    statistics = nextmv.Statistics(
        run=run,
        result=nextmv.ResultStatistics(
            custom={
                "depth": model.get_depth(),
                "feature_importances_": model.feature_importances_.tolist(),
                "n_leaves": int(model.get_n_leaves()),
                "score": model.score(X, y, sample_weight),
            },
        ),
        series_data=nextmv.SeriesData(),
    )

    if sample_weight is not None:
        statistics.result.custom["sample_weight"] = sample_weight

    return statistics