EPFL ML Project 1 - detecting the Higgs Boson.

If you're using Windows, some steps might differ.

1. Python 3

2. PIP

3. Numpy

1. Install numpy

pip3 install numpy

2. Change current directory to the project's folder

cd higgs

3. Create datasets/ folder at the root of the project folder and copy there the train.csv and test.csv files.
4. Run the following command:

python3 run.py

• run.py - this file contains the final pipeline, which is a bag of neural networks. It is a simplified version of neural_bagging.py.
• run_nn.py - a pipeline that trains a single neural network. Depending on the options specified inside the file, it either runs it on a validation set and prints the result, or on test set and produces predition.csv
• neural_bagging.py - a pipeline that trains several neural networks and combines them with majority voting. Depending on the options specified inside the file, it either runs it on a validation set and prints the result, or on test set and produces predition.csv
• simple_net.py - A neural network implementation. See the file for detailed parameters behind it. A typical usage involes 2 calls - fit() to train a network and predict() to get predictions on some set.
• run_linear.py - Used to run pipelines on regressions.
• majority_combinator.py - function that combines several predictions into one using majority voting.
• utils.py - Utility for checking intesresction of two predictions (what they both get right, wrong, etc)
• helpers.py - Various helper functions.
• implementations.py - Implementaions of basic methods.
• featurization.py - Featurization pipeline.
• hyperparapemeter_nn_grid_search.py - Grid search utility to pick the best hyperparameters.
• DecisionTree.py - decision tree.