This repository explores clustering techniques applied to astronomical star datasets, with a particular emphasis on evolutionary and unsupervised clustering approaches using Python.
The goal is to analyze and visualize natural groupings of stars based on their features using different clustering strategies such as k-means and DBSCAN.
Unsupervised clustering aims to discover structure in datasets without ground-truth labels. In this project:
- k-means clustering assigns each star to one of k clusters based on feature similarity.
- DBSCAN detects clusters based on density and neighborhood relationships.
These methods help identify natural groupings in star datasets that may correspond to physical or observational characteristics.
The notebooks demonstrate how clustering algorithms can be applied, visualized, and evaluated in a scientific context.
To run the code or notebooks locally, you’ll typically need:
- Python 3.x
- Jupyter Notebook
- Libraries such as:
pandasnumpymatplotlibscikit-learn
You can install common dependencies with:
pip install numpy pandas scikit-learn matplotlib notebook