AstroFusion: A GAN-Augmented Approach for Exoplanet Detection

Humanity's exploration of the cosmos spans billions of years, prompting contemplation of our position in the universe. This study advocates for proactive communication with potential intelligent species, driven by uncertainty about their existence. The focus is on exploring Generative Adversarial Networks (GANs) to enhance exoplanet detection through data augmentation. Utilizing the extensive Kepler dataset, the study evaluates the accuracy of conventional machine learning techniques: Logistic Regression, Decision Trees, and Random Forests. (RF)-trained on original data. Subsequently, it assesses the effectiveness of models trained solely on GAN-generated synthetic data and compares their performance with that of models trained on original data. Lastly, it merges original and synthetic data to create a composite dataset, evaluating the ML model's accuracy. The results indicate that models trained solely on synthetic data achieve comparable accuracy to those trained on original data. Furthermore, the combined dataset yields improved accuracy-87%, 95%, and 96% in Logistic Regression, Decision Tree, and RF, respectively-highlighting the potential of GAN-based data augmentation in enhancing exoplanet detection. This study advances space exploration by integrating technology with astronomical research, facilitating more precise and reliable exoplanet detection methods. Leveraging GANs for data augmentation promises novel insights and discoveries about the universe.