Akademik Veri Yönetim Sistemi
The 2025 conference on Big Data from Space (BiDS'25), Riga, Letonya, 29 Eylül - 03 Ekim 2025, ss.173-176, (Özet Bildiri)
Asteroid collision prediction plays a pivotal role in planetary defence by enabling proactive risk mitigation and informed strategic planning. To address this challenge, we developed a comprehensive framework that integrates historical fireball data and contemporary orbital parameters from NASA’s datasets. By deriving shared physical features and applying unsupervised clustering, our system identifies patterns in potential impact scenarios. We further incorporate supervised learning to categorize asteroids based on their threat level. To enhance accessibility and interpretation, the framework includes multi-dimensional visualizations of orbital dynamics and an interactive web application that represents asteroid trajectories in both two and three dimensions. This simulation platform serves both scientific and educational purposes, offering a rich interface for exploring asteroid behaviour. This study demonstrates the potential of combining machine learning, astrophysical modelling, and data visualization to support planetary safety initiatives.