MLPR Project: Pulsar detection

Team members

• Francesco Scalera (s292432)
• Riccardo Sepe (s287760)

Project structure

The core of our project is the classifiers folder: it contains the classes that model the classifiers we used. The base abstract class ClassifierClass is contained in Classifier.py: its subclasses can be constructed by providing the training data and labels and optionally some **kwargs that will vary based on which is the subclass. It exposes three abstract methods: train_model(), classify() and get_scores(). The latter is used to return the classifiers scores generated inside classify(), so it is not directly necessary for classification, but it's useful for our purpose of evaluating the performance of our model itself in the Optimal Bayes Decision (DCF computation) framework. The preprocessing folder contains code useful for the pre-processing steps we considered and the utils folder contains some utility functions grouped by purpose: matrix_utils, plot_utils and misc_utils. The data, results and simulations folders contain respectively the data, the results in terms of DCFs for each classifier and for each possible configuration and various collections of .npy files containing the values for the plots. In file main.py is loaded the data and are called the tuning functions and simulations functions.

Project requirements

The project requirements are gathered in the requirements.txt file. They are:

• distinctipy: used to generate distinguishable colors for the plots
• matplotlib: used to produce all the plots. As an additional requirement there must be a Latex compiler on the machine running the code to produce Latex labels
• numpy: for numerical computations
• prettytable: to produce human-readable tables with all the results
• scipy: for numerical computations

Credits

R. J. Lyon, B. W. Stappers, S. Cooper, J. M. Brooke, J. D. Knowles, Fifty Years of Pulsar Candidate Selection: From simple filters to a new principled real-time classification approach, Monthly Notices of the Royal Astronomical Society 459 (1), 1104-1123, DOI: 10.1093/mnras/stw656

R. J. Lyon, HTRU2, DOI: 10.6084/m9.figshare.3080389.v1.