Binary classifier metrics for optimizing HEP event selection

I discuss the choice of evaluation metrics for binary classifiers in High Energy Physics (HEP) event selection and I point out that the Area Under the ROC Curve (AUC) is of limited relevance in this context, after discussing its use in other domains. I propose new metrics based on Fisher information...

Descripción completa

Detalles Bibliográficos
Autor principal: Valassi, Andrea
Lenguaje:eng
Publicado: 2019
Materias:
Computing and Computers
Particle Physics - Experiment
Acceso en línea:https://dx.doi.org/10.1051/epjconf/201921406004
http://cds.cern.ch/record/2699582
_version_ 1780964498506514432
author Valassi, Andrea
author_facet Valassi, Andrea
author_sort Valassi, Andrea
collection CERN
description I discuss the choice of evaluation metrics for binary classifiers in High Energy Physics (HEP) event selection and I point out that the Area Under the ROC Curve (AUC) is of limited relevance in this context, after discussing its use in other domains. I propose new metrics based on Fisher information, which can be used for both the evaluation and training of HEP event selection algorithms in statistically limited measurements of a parameter.
id oai-inspirehep.net-1761270
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2019
record_format invenio
spelling oai-inspirehep.net-17612702022-08-10T12:21:25Zdoi:10.1051/epjconf/201921406004http://cds.cern.ch/record/2699582engValassi, AndreaBinary classifier metrics for optimizing HEP event selectionComputing and ComputersParticle Physics - ExperimentI discuss the choice of evaluation metrics for binary classifiers in High Energy Physics (HEP) event selection and I point out that the Area Under the ROC Curve (AUC) is of limited relevance in this context, after discussing its use in other domains. I propose new metrics based on Fisher information, which can be used for both the evaluation and training of HEP event selection algorithms in statistically limited measurements of a parameter.oai:inspirehep.net:17612702019
spellingShingle Computing and Computers
Particle Physics - Experiment
Valassi, Andrea
Binary classifier metrics for optimizing HEP event selection
title Binary classifier metrics for optimizing HEP event selection
title_full Binary classifier metrics for optimizing HEP event selection
title_fullStr Binary classifier metrics for optimizing HEP event selection
title_full_unstemmed Binary classifier metrics for optimizing HEP event selection
title_short Binary classifier metrics for optimizing HEP event selection
title_sort binary classifier metrics for optimizing hep event selection
topic Computing and Computers
Particle Physics - Experiment
url https://dx.doi.org/10.1051/epjconf/201921406004
http://cds.cern.ch/record/2699582
work_keys_str_mv AT valassiandrea binaryclassifiermetricsforoptimizinghepeventselection

Ejemplares similares