EUSC: A clustering-based surrogate model to accelerate evolutionary undersampling in imbalanced classification

Learning from imbalanced datasets is highly demanded in real-world applications and a challenge for standard classifiers that tend to be biased towards the classes with majority of examples. Undersampling approaches reduce size class balance distributions. Evolutionary-based are prominent, treating undersampling as binary optimisation problem determines which examples removed. However, their utilisation limited small due fitness evaluation costs. This work proposes two-stage clustering-based surrogate model enables evolutionary compute values faster. The main novelty lies development based on meaning (phenotype) rather than representation (genotype). We conduct an 44 datasets, showing comparison original undersampling, we can save up 83% runtime without significantly deteriorating classification performance.