StereoMeanCombiner#

class ctapipe.reco.stereo_combination.StereoMeanCombiner(*args: t.Any, **kwargs: t.Any)[source]#

Calculate array-event prediction as (weighted) mean of telescope-wise predictions.

Attributes Summary

`log_target`	If true, calculate exp(mean(log(values))).
`weights`	What kind of weights to use.

Methods Summary

`__call__`(event)	Calculate the mean prediction for a single array event.
`predict_table`(mono_predictions)	Calculates the (array-)event-wise mean.

Attributes Documentation

weights#: What kind of weights to use. Options: none, intensity, aspect-weighted-intensity, containment-weighted-intensity. containment-weighted-intensity down-weights truncated (leaky) images via (intensity * (1 - width / length))**2 * (1 - leakage_intensity_width_2)**4 and is the most robust option at high energies where image truncation dominates.

Methods Documentation

__call__(event: ArrayEventContainer) → None[source]#: Calculate the mean prediction for a single array event.

predict_table(mono_predictions: Table) → Table[source]#: Calculates the (array-)event-wise mean. Telescope events, that are nan, get discarded. This means you might end up with less events if all telescope predictions of a shower are invalid.