Aim
To develop an improved credit-scoring and text-classification model by combining Bi-LSTM, advanced transformer architectures, and ensemble machine-learning methods for superior accuracy, robustness, and fairness when compared with existing hybrid systems.