get_evt_model_output#

protopipe.mva.get_evt_model_output(data_dict, weight_name=None, reco_energy_label='reco_energy', model_output_name='score_img', model_output_name_evt='score')[source]#

Returns DataStore with reco energy + score/target columns of model at the level-event.

Parameters
data: `~pandas.DataFrame`

Data frame with at least (obs_id, evt_id), image score/proba and label

weight_name: `str`

Variable name in data frame to weight events with

reco_energy_label: `list`, optional

List of variables to keep in resulting data frame

model_output_name: `str`, optional

Name of model output (image level)

model_output_name: `str`, optional

Name of averaged model output (event level)

Returns
data: ~pandas.DataFrame

Data frame

Warning

Need optimisation, too much loops.