Team Members: Anshika, Niegil, Sakthisree, Vishnu
PDF also available in the repo.
Link to the Notebook Link Submission for Higgs Boson EDA, Baseline & Ensemble: Click here to see notebook
Link to Notebook with VIME implementation of self supervised learning in the tabular domain: Click here to see notebook
Link to Report and Video: Click here to open Drive Folder
The discovery of Higgs particle was announced on 4th July 2012. In 2013, Nobel Prize was conferred upon two scientists, Francois Englert and Peter Higgs for their contribution towards its discovery. A characteristic property of Higgs Boson is its decay into other particles through different processes. At the ATLAS detector at CERN, very high energy protons are accelerated in a circular trajectory in both directions thus colliding with themselves and resulting in hundreds of particles per second.
These events are categorized as either background or signal events. The background events consist of decay of particles that have already been discovered in previous experiments. The signal events are the decay of exotic particles: a region in feature space which is not explained by the background processes. The significance of these signal events is analyzed using different statistical tests.
If the probability that the event has not been produced by a background process is well below a threshold, a new particle is considered to have been discovered. The ATLAS experiment observed a signal of the Higgs Boson decaying into two tau particles, although it was buried in significant amount of noise.
The goal of the Challenge is to improve the classification procedure that produces the selection region.The objective is a function of the weights of selected events. Prefix-less variables EventId, Weight and Label have a special role and should not be used as input to the classifier.
The variables prefixed with PRI (for PRImitives) are “raw” quantities about the bunch collision as measured by the detector, essentially the momenta of particles. Variables prefixed with DER (for DERived) are quantities computed from the primitive features.
- Data Loading
- Data Pre processing - Removing Highly correlated features, log transformations, and SMOTE upsampling
- Baseline Logistic Regression - 71% Accuracy
- Decision Tree Classifier - 78% Accuracy
- Ensemble - Random Forest Classifier - 84% Accuracy
- Ensemble - Bagging Classigier - 84% Accuracy
- VIME Self-supervised implementation - 54% Accuracy