Machine-learned exclusion limits without binning
Abstract Machine-learned likelihoods (MLL) combines machine-learning classification techniques with likelihood-based inference tests to estimate the experimental sensitivity of high-dimensional data sets. We extend the MLL method by including kernel density estimators (KDE) to avoid binning the clas...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2023-12-01
|
| Series: | European Physical Journal C: Particles and Fields |
| Online Access: | https://doi.org/10.1140/epjc/s10052-023-12314-z |
| _version_ | 1797233306019823616 |
|---|---|
| author | Ernesto Arganda Andres D. Perez Martín de los Rios Rosa María Sandá Seoane |
| author_facet | Ernesto Arganda Andres D. Perez Martín de los Rios Rosa María Sandá Seoane |
| author_sort | Ernesto Arganda |
| collection | DOAJ |
| description | Abstract Machine-learned likelihoods (MLL) combines machine-learning classification techniques with likelihood-based inference tests to estimate the experimental sensitivity of high-dimensional data sets. We extend the MLL method by including kernel density estimators (KDE) to avoid binning the classifier output to extract the resulting one-dimensional signal and background probability density functions. We first test our method on toy models generated with multivariate Gaussian distributions, where the true probability distribution functions are known. Later, we apply the method to two cases of interest at the LHC: a search for exotic Higgs bosons, and a $$Z'$$ Z ′ boson decaying into lepton pairs. In contrast to physical-based quantities, the typical fluctuations of the ML outputs give non-smooth probability distributions for pure-signal and pure-background samples. The non-smoothness is propagated into the density estimation due to the good performance and flexibility of the KDE method. We study its impact on the final significance computation, and we compare the results using the average of several independent ML output realizations, which allows us to obtain smoother distributions. We conclude that the significance estimation turns out to be not sensible to this issue. |
| first_indexed | 2024-03-08T19:44:11Z |
| format | Article |
| id | doaj.art-27abc3497fb14803b94a5c3ee3930c74 |
| institution | Directory Open Access Journal |
| issn | 1434-6052 |
| language | English |
| last_indexed | 2024-04-24T16:14:04Z |
| publishDate | 2023-12-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | European Physical Journal C: Particles and Fields |
| spelling | doaj.art-27abc3497fb14803b94a5c3ee3930c742024-03-31T11:30:59ZengSpringerOpenEuropean Physical Journal C: Particles and Fields1434-60522023-12-01831211410.1140/epjc/s10052-023-12314-zMachine-learned exclusion limits without binningErnesto Arganda0Andres D. Perez1Martín de los Rios2Rosa María Sandá Seoane3Departamento de Física Teórica, Universidad Autónoma de MadridDepartamento de Física Teórica, Universidad Autónoma de MadridDepartamento de Física Teórica, Universidad Autónoma de MadridInstituto de Física Teórica UAM-CSICAbstract Machine-learned likelihoods (MLL) combines machine-learning classification techniques with likelihood-based inference tests to estimate the experimental sensitivity of high-dimensional data sets. We extend the MLL method by including kernel density estimators (KDE) to avoid binning the classifier output to extract the resulting one-dimensional signal and background probability density functions. We first test our method on toy models generated with multivariate Gaussian distributions, where the true probability distribution functions are known. Later, we apply the method to two cases of interest at the LHC: a search for exotic Higgs bosons, and a $$Z'$$ Z ′ boson decaying into lepton pairs. In contrast to physical-based quantities, the typical fluctuations of the ML outputs give non-smooth probability distributions for pure-signal and pure-background samples. The non-smoothness is propagated into the density estimation due to the good performance and flexibility of the KDE method. We study its impact on the final significance computation, and we compare the results using the average of several independent ML output realizations, which allows us to obtain smoother distributions. We conclude that the significance estimation turns out to be not sensible to this issue.https://doi.org/10.1140/epjc/s10052-023-12314-z |
| spellingShingle | Ernesto Arganda Andres D. Perez Martín de los Rios Rosa María Sandá Seoane Machine-learned exclusion limits without binning European Physical Journal C: Particles and Fields |
| title | Machine-learned exclusion limits without binning |
| title_full | Machine-learned exclusion limits without binning |
| title_fullStr | Machine-learned exclusion limits without binning |
| title_full_unstemmed | Machine-learned exclusion limits without binning |
| title_short | Machine-learned exclusion limits without binning |
| title_sort | machine learned exclusion limits without binning |
| url | https://doi.org/10.1140/epjc/s10052-023-12314-z |
| work_keys_str_mv | AT ernestoarganda machinelearnedexclusionlimitswithoutbinning AT andresdperez machinelearnedexclusionlimitswithoutbinning AT martindelosrios machinelearnedexclusionlimitswithoutbinning AT rosamariasandaseoane machinelearnedexclusionlimitswithoutbinning |