feos.eos.estimator.Estimator¶

class feos.eos.estimator.Estimator(data, weights, losses)¶

A collection of `DataSet`s that can be used to compute metrics for experimental data.

Parameters:

data (List[DataSet]) – The properties and experimental data points to add to the estimator.
weights (List[float]) – The weight of each property. When computing the cost function, the weights are normalized (sum of weights equals unity).
losses (List[Loss]) – The loss functions for each property.

Return type:

Methods

`__init__`()
`cost`(eos)	Compute the cost function for each `DataSet`.
`mean_absolute_relative_difference`(eos)	Return the mean absolute relative difference for each `DataSet`.
`predict`(eos)	Return the properties as computed by the equation of state for each DataSet.
`relative_difference`(eos)	Return the relative difference between experimental data and prediction of the equation of state for each `DataSet`.

Attributes

datasets

Return the stored ``DataSet``s.