CERN Accelerating science

MatrixNet: Using a new Multivariate Technique in High Energy Physics

Date published: 
Sunday, 1 September, 2013
Document type: 
Summer student report
Author(s): 
V. Doneva
This project focuses on testing and developing algorithms for multivariate data analysis, that separate signal processes from abundant backgrounds and on helping with organizing and filtering colossal amounts of raw data, gathered from the Large Hadron Collider beauty (LHCb) experiment, to find extremely rare events of interest. Moreover, working on this project also meant trying to apply new, faster method to take the place of systems that are now used at CERN to pare down the relevant data, but require relatively extensive processing and analysis to determine relevance and usefulness.