Time calibration of barrel TOF system at BESIII

Purpose In the BESIII experiment,a conventional least square method is currently used to calibrate the time-of-flight(TOF)system,aiming to achieve optimal time resolution.However,it is intricate for conventional methods to handle electronics saturation and boundary effect which affect the time resolution.Methods A novel method based on machine learning techniques for time calibration of barrel TOF system is implemented in this paper.Results and conclusion With the application of boosted decision tree(BDT)algorithm,the overall time resolution of barrel TOF is found to be 64 ps.Compared with conventional method,a notable improvement in time resolution is achieved in the boundary range of the scintillator bar using BDT method.