Higgs data mining project (2015)
This project was done as a part of introduction to data mining and machine learning course at Tel Aviv University.
Data set is provided by Kaggle challenge: https://www.kaggle.com/c/higgs-boson
The goal is to classify Boson Higgs events to signal (real event) and noise given measured attributes.
The full data can be found at Kaggle link (only sample files appear at the repository).
The project includes:
- project.py:
- Feature Selection (T-Test, FDR test)
- KNN Classifier (68% accuracy)
- Decision Tree Classifier (81% accuracy)
- Random Forest Classifier (80% accuracy)
- ANN.py:
- Neural Network Classifier (82% accuracy)