Machine Learning Project 1
Asli Yorusun: asli.yorusun@epfl.ch
Erdem Bocugoz: erdem.bocugoz@epfl.ch
Serif Soner Serbest: serif.serbest@epfl.ch
In this project we predict CERNs simulated particle collision events as either a Higgs Boson signal or background noise as of binary classification, which is a Kaggle
We ranked 28th in Kaggle LeaderBoard among 211 teams,with our score: 0.83224.
To get the exact results run the "run.py" file.
labels,features,data,id_ = load_csv_data(data_path, sub_sample=False)
3 subsest according to Jet category = categorize_data(prediction, data)
Cleaned and standartized features = process_data(features) build_model_data(prediction, x)
Polynomially Expanded Features = build_poly(features, degree) :
weight, loss = reg_logistic_regression(labels,feature_model,lambda,max iteration, gamma)
weight, loss = ridge_regression(labels, feature_model, lambda)
prediction = predict_labels(weight,feature_model)
labels = decategorize_prediction(row_size, label1, label2, label3, indices1, indices2, indices3)
create_csv_submission(test_id_,labels,"submission.csv")