Imbalanced classification applied to asteroid resonant dynamics

Introduction: Machine learning (ML) applications for studying asteroid resonant dynamics are a relatively new field of study. Results from several different approaches currently available asteroids interacting with the z 2 , 1 M1:2, and ν 6 resonances. However, one challenge when using ML to databases produced by these studies is that there often severe imbalance ratio between number in librating orbits rest asteroidal population. This can be as high 1:270, which impact performance classical algorithms, were not designed such imbalances. Methods: Various techniques have been recently developed address this problem, including cost-sensitive strategies, methods oversample minority class, undersample majority one, or combinations both. Here, we investigate most effective improving algorithms known databases. Results: Cost-sensitive either improved had affect outcome should always used, possible. The showed best studied SMOTE oversampling plus Tomek undersampling, oversampling, Random undersampling. Discussion: Testing first could save significant time efforts future imbalanced