The European Organization for Nuclear Research (CERN) Machine Learning Research
A short summary of this research could be found at Research_Summary.pdf and Research Poster.pdf
▪ Implemented a set of analysis scripts in Python and Bash under ROOT framework, including preprocessing events data
and building the training/testing pipelines of different Decision Tree models (Boosted Gradient, Adaptive, with Fisher
discriminant and decorrelation).
▪ Improved signal efficiency by 7% and background rejection by 12.5%; replaced the old events classification model,
which was based on threshold cuts of discriminant variables.