Weighted Harmonic Mean¶

class pumas.aggregation.weighted_harmonic_mean.WeightedHarmonicMeanAggregation(params: Dict[str, Any] | None = None)[source]

Computes the weighted harmonic mean of a set of values with corresponding weights.

\[H = \frac{\sum_{i=1}^{n} w_i}{\sum_{i=1}^{n} \frac{w_i}{x_i}}\]

Where:

\(H\) is the weighted harmonic mean
\(x_i\) are the values
\(w_i\) are the corresponding weights
\(n\) is the number of observations

Example: >>> from pumas.aggregation import aggregation_catalogue

>>> aggregator_class = aggregation_catalogue.get("harmonic_mean")

>>> aggregator = aggregator_class()

>>> values = [1.0, 2.0, 3.0]
>>> weights = [0.2, 0.3, 0.5]
>>> result = aggregator.compute_numeric(values=values, weights=weights)
>>> print(f"{result:.2f}")
1.94

>>> result = aggregator(values=values, weights=weights) # Same as compute_numeric
>>> print(f"{result:.2f}")
1.94

>>> from uncertainties import ufloat
>>> values = [ufloat(1.0, 0.1), ufloat(2.0, 0.2), ufloat(3.0, 0.3)]
>>> weights = [0.2, 0.3, 0.5]
>>> result = aggregator.compute_ufloat(values=values, weights=weights)
>>> print(result)
1.94+/-0.11

compute_numeric(values: List[float | None], weights: List[float | None] | None = None) → float[source]

Compute the weighted harmonic mean for uncertain float input values.

Parameters:

values (List[float]) – The list of uncertain float values to be aggregated.
weights (Optional[List[float]]) – The list of weights corresponding to each value. If None, equal weights are assumed.

Returns:

The computed weighted harmonic mean with uncertainty.

Return type:

UFloat

compute_ufloat(values: List[AffineScalarFunc | None], weights: List[float | None] | None = None) → AffineScalarFunc[source]