This paper examines the classification performance of artificial immune systems on the one hand and machine learning and neural networks on the other hand on the problem of forecasting credit ratings of firms. The problem is realized as a two-class problem, for investment and non-investment rating grades. The dataset is usually imbalanced in credit rating predictions. We address the issue by over-sampling the minority class in the training dataset. The experimental results show that this approach leads to significantly higher classification accuracy. Additionally, the use of the ensembles of classifiers makes the prediction even more accurate.