Prediction of RPC efficiency using machine learning techniques

In this work, K-NN classification and regression models are used to estimate the efficiency for the Resistive Plate Chambers (RPC) of the Compact Muon Solenoid (CMS), one of the Large Hadron Collider (LHC) experiments located at CERN between the French and Swiss border. Measurable detector quantities are used as input parameters to estimate the efficiency. Ten datasets taken by the CMS experiment during 2018 are used. The first set (80% of the data) is used to perform a training for the K-NN model, and the second set (20%) is used to make an efficiency prediction using the training performed with the initial set. Finally, a comparison between the real value of the efficiency and the value of the prediction is performed. The results obtained are promising.