A competition on Kaggle to identify a rare decay phenomenon.
This is a project done with Chris. It’s a competition on Kaggle this year called "Flavours of Physics Finding τ → μμμ", which goal is to find a phenomenon that is not already known to exist – charged lepton flavour violation – thereby helping to establish “new physics”.
A list of collision events and their properties is given. The mission is to predict whether a τ → 3μ decay happened in this collision. This τ → 3μ is currently assumed by scientists not to happen, and the goal of this competition is to discover τ → 3μ happening more frequently than scientists currently can understand. Basically this can be treated as a two-class classification problem. The dataset is labelled (the label ‘signal’ being ‘1’ for signal events, ‘0’ for background events) to train the classifier.
A Random Forest and a Gradient Boost classifiers are trained with the dataset. To attain high performance, linear combination of the output scores of an ensemble of classifiers is used. In the end, our result score is up to 0.991063 and ranked within top 10% (19th/673).
A snapshot of my Kaggle account
