A set of files examining the performance of different classifiers on the MAGIC telescope simulated data.

Files:

magic.py - computes and plots ROC curves for a variety of classifiers.

magic_nn.py - Trains a feedforward neural network on the MAGIC dataset using PyBrain. Saves trained network to file magic_nn.xml. Plots cost function vs training round.

magic2_nn.py - Similar to magic_nn.py, but uses different PyBrain commands. Saves trained network to file magic2_nn.xml. Plots training accuracy vs training round.