processing.agg_predict.agg_predict

processing.agg_predict.agg_predict(predictions, weights=None)

Aggregates multiple prediction columns into a single combined prediction series.

The combination is a weighted sum of the prediction columns. If no weights are provided, a default weighting scheme based on specific predefined columns is used.

Parameters

Name	Type	Description	Default
predictions	pd.DataFrame	DataFrame containing the prediction columns.	required
weights	Optional[Union[Dict[str, float], List[float], np.ndarray]]	Dictionary mapping column names to their weights, or a list/array of weights corresponding to the order of columns in `predictions`. If None, defaults to summing all columns (weight=1.0 for each column).	`None`

Returns

Name	Type	Description
	pd.Series	pd.Series: A Series containing the aggregated values.

Raises

Name	Type	Description
	ValueError	If a column specified in weights (or default weights) is missing from predictions.
	ValueError	If weights is a list/array and its length does not match the number of columns in predictions.

Examples

>>> from spotforecast2_safe.processing import agg_predict
>>> import pandas as pd
>>> df = pd.DataFrame({"A": [1, 2], "B": [3, 4]})
>>> agg_predict(df, weights={"A": 1.0, "B": -1.0})
0   -2.0
1   -2.0
dtype: float64
>>> agg_predict(df, weights=[0.5, 2.0])
0    6.5
1    9.0
dtype: float64