Pulsar-Detection

Using Machine Learning for Pulsar Detection

Final Report (PDF)

Final report can be downloaded here: Report

Task

Pulsars are kind of Neutron stars and considerably interesting for scientific research. As this exceptional kind of star produces radio emissions detectable here on Earth, machine learning tools can be used to label pulsar candidates to facilitate rapid analysis automatically. Classification systems are widely implemented, considering the candidate data sets as binary classification problems.

Candidate information

Each candidate is described by eight continuous variables and a single class variable. The first four are simple statistics obtained from the integrated pulse profile (folded profile). This is an array of continuous variables that describe a longitude-resolved version of the signal that has been averaged in both time and frequency. The remaining four variables are similarly obtained from the DM-SNR curve.

1. Mean of the integrated profile.
2. The standard deviation of the integrated profile.
3. Excess kurtosis of the integrated profile.
4. The skewness of the integrated profile.
5. Mean of the DM-SNR curve.
6. The standard deviation of the DM-SNR curve.
7. Excess kurtosis of the DM-SNR curve.
8. The skewness of the DM-SNR curve.
9. Class

Classification task

Given the features above, our final task is to determine whether each of the samples is a Pulsar candidate or not. So, we are faced with a binary classification problem, and we need to build different classifiers, analyze them, and compare their performance and cost for various applications.

What has been done

FEATURE AND CLASS ANALYSIS

• CLASS DISTRIBUTION
• FEATURE DISTRIBUTION
• CORRELATION ANALYSIS

CLASSIFICATION

• APPLICATIONS AND CROSS-VALIDATION
• MVG CLASSIFIERS
• LOGISTIC REGRESSION
• QUADRATIC LOGISTIC REGRESSION
• LINEAR SVM
• KERNEL SVM
• GAUSSIAN MIXTURE MODEL (GMM)