Pulsar Classification

This personal project consists of Python scripts to compare standard sci-kit learn classification algorithms and keras/tensorflow neural nets. We use HTRU2 data from the UCI archive, namely statistical data on suspected pulsars, to classify radio telescope observations as Noise (0) or Pulsar (1)

Classification algorithms, gridsearch parameter tuning

Logistic regression, Decision Trees, Random Forest, KNeighbors, Naive Bayes, Artificial Neural Net (keras/tensorflow)

GitHub Repo

Original Dataset

Special note, I have a similar project for regression algorithms, including PCA dimensionality reduction, designed to predict the continuous variable of superconductor critical temperatures:

GitHub Repo for Superconductor Temps

Previous
Previous

Climate Change Analysis (2021)