The article introduces a new decision tree method that helps balance out imbalanced data sources, improving the efficiency of decision-making models. By using a sampling strategy, the method mimics human learning and shows comparable performance to other advanced methods.