Abstract
The common approach for constructing a classifier for particle selection assumes reasonable consistency between train data samples and the target data sample used for the particular analysis. However, train and target data may have very different properties, like energy spectra for signal and background contributions. We propose a new method based on an ensemble of pre-trained classifiers, each trained of an exclusive subset, a data basket, of the total dataset. Appropriate separate adjustment of separation thresholds for every basket classifier allows to dynamically adjust the combined classifier and make optimal prediction for data with differing properties without re-training of the classifier. The approach is illustrated with a toy example. A quality dependency on the number of used data baskets is also presented.
Highlights
The common approach to event selection to boost the signal of background ratio in particle physics analysis assumes calibration and/or MC samples to train corresponding classifiers
If in some parts of the spectrum there is a clear dominance of one type of analyzed objects, the classifier would tend to assign this type to new objects only because there is such a big contribution in the training sample
Further adjustments to real spectra of particular data analyses does not require re-training if using a priori knowledge of shapes of the target data sets
Summary
The common approach to event selection to boost the signal of background ratio in particle physics analysis assumes calibration and/or MC samples to train corresponding classifiers. Such a pre-trained classifier may be used in physical analysis later [1, 2]. This approach allows advanced training and validation of common purpose classifiers beforehand, and applying them to different physics analyses in a smooth and uniform way. If in some parts of the spectrum there is a clear dominance of one type of analyzed objects, the classifier would tend to assign this type to new objects only because there is such a big contribution in the training sample
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
