Class imbalance can significantly affect the accuracy and reliability of predictive models. When one class dominates the dataset, models tend to be biased towards that class, leading to poor model performance on the minority class. This is particularly problematic in catalysis research, where discovering rare but effective catalysts is crucial.
