19–23 Jun 2023
Hofburg Palace & Online
Europe/Vienna timezone

A Novel Radioxenon Event Classification Method Based on CatBoost

P2.1-553
21 Jun 2023, 09:00
1h
Wintergarten

Wintergarten

Board: 13
E-poster T2.1 Characterization of Treaty-Relevant Events Lightning talks: P2.1, P2.3, P4.4

Speaker

Mr Yungang Zhao (CTBT Beijing National Data Centre and Beijing Radionuclide Laboratory)

Description

Four technologies are used for compliance verification of the comprehensive Nuclear-Test Ban Treaty (CTBT), hydroacoustic, infrasound, seismic and radionuclide monitoring. Forty radionuclide stations will have the radioxenon monitoring capability. For the last few years, a lot of research have been done in order to discriminate radioxenon derived from an explosion and anthropogenic sources. The explosion data are rare, for quite few nuclear test were done and little radioxenon was released from the underground. The civil facilities also emit mounts of radioxenon and the stations near those facilities will have a high radioxenon background. It is difficult to discriminate the background events and background plus explosion events. In this work, radionxenon monitoring data sets were modelled based on International Monitoring System xenon monitoring data over the past decade and the xenon ratios emitted by nuclear explosion. Dozens of features, including the four radionxenon concentrations, flags on xenon detected, xenon ratios, were analysed. We used CatBoost algorithm to establish a two classification model.

Promotional text

Radionxenon monitoring data sets were modelled based IMS xenon monitoring data over the past decade and the xenon ratios emitted by nuclear explosion. CatBoost algorithm is used to establish a two classification model based on dozens of features.

E-mail [email protected]
Oral preference format online live

Primary author

Mr Yungang Zhao (CTBT Beijing National Data Centre and Beijing Radionuclide Laboratory)

Co-authors

Ruiying Li Mr Shilian Wang (CTBT Beijing National Data Centre and Beijing Radionuclide Laboratory) Qi Li (CTBT Beijing National Data Centre and Beijing Radionuclide Laboratory) Mr Xinjun Zhang Yuanqing Fan Huaimao Jia (CTBT Beijing National Data Centre and Beijing Radionuclide Laboratory) Ruiqin Zhang (CTBT Beijing National Data Centre and Beijing Radionuclide Laboratory)

Presentation materials