https://github.com/prajakta1321/exoplanet-atmospheric-characterization-1
A machine learning project to classify exoplanets using light curve image data. Developed as part of the ML4SCI GSoC 2025 Test Task. Includes data processing, CNN-based model, and full report.
https://github.com/prajakta1321/exoplanet-atmospheric-characterization-1
classification colab-notebook dbscan gsoc-2025 machine-learning-algorithms matplotlib-python ml numpy open-source pca-analysis python3 seaborn
Last synced: 2 months ago
JSON representation
A machine learning project to classify exoplanets using light curve image data. Developed as part of the ML4SCI GSoC 2025 Test Task. Includes data processing, CNN-based model, and full report.
- Host: GitHub
- URL: https://github.com/prajakta1321/exoplanet-atmospheric-characterization-1
- Owner: prajakta1321
- Created: 2025-04-08T07:40:27.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2025-04-08T07:46:37.000Z (2 months ago)
- Last Synced: 2025-04-08T08:40:25.664Z (2 months ago)
- Topics: classification, colab-notebook, dbscan, gsoc-2025, machine-learning-algorithms, matplotlib-python, ml, numpy, open-source, pca-analysis, python3, seaborn
- Language: Jupyter Notebook
- Homepage:
- Size: 0 Bytes
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Exoplanet-atmospheric-characterization-1
# 🌌 Exoplanet Classification using Machine Learning
This project was developed as part of the GSoC 2025 Test Task for the ML4SCI (Machine Learning for Science) organization. The goal is to build a machine learning pipeline to classify exoplanets based on their light curve image data using supervised learning techniques.
## 📌 Project Objective
To explore, build, and evaluate an efficient ML model that can accurately classify whether a given celestial body is an exoplanet or not, using image-based data derived from light curves.
## 🛠️ Tools & Technologies Used
- Python
- Pandas & NumPy
- Matplotlib & Seaborn
- TensorFlow / Keras
- Scikit-learn
- Google Colab
## 📂 Project Structure as follows
## 🚀 Steps Performed
1. **Data Preparation**:
- Converted light curve CSVs into grayscale image data.
- Stored images in a local folder for manual upload.
2. **Image Preprocessing**:
- Resized images, normalized pixel values, and converted to NumPy arrays.
3. **Model Building**:
- Used Convolutional Neural Networks (CNNs) for classification.
- Achieved a satisfactory training and validation accuracy.
4. **Model Saving**:
- Trained model saved as `model.h5` for reuse and deployment.
5. **Report Generation**:
- Includes insights, challenges, learnings, and future scope.
## 📊 Results
- Achieved high accuracy on the validation dataset.
- Demonstrated effective use of image-based ML in scientific tasks.
## 🧠 What I Learned
- Converting CSV-based signal data to image form.
- Building and training CNNs on image datasets.
- Handling limited data using augmentation techniques.
- Structuring and documenting ML projects.
## 🤝 Contributions
If accepted, I look forward to contributing further to the ML4SCI initiative by refining this pipeline and extending it to more complex datasets.
---