Abstract: Back Propagation (BP) is widely used in credit rating, but it has such shortcomings as slow convergent speed and easy convergence to the local minimum points. A credit rating model that combined ant colony algorithm with neural network was proposed. It not only has the extensive mapping ability of neural network, but also has the advantages of high efficiency, rapid global convergence and distributed computation of ant system. The experimental result indicates good performance can be obtained by neural network based on ant colony algorithm in application of the credit rating.

Key words: ant colony algorithm, neural network, credit rating, BP algorithm

摘要： BP算法在资信评估中应用较为广泛，但有收敛速度慢、易于陷入局部极小点的缺点。提出一种新的企业资信评估模型，该模型将蚁群算法和神经网络结合起来，使其既具有神经网络的广泛映射能力，又有蚁群算法带来的高效率，全局收敛，分布式计算等特点。实验表明,基于蚁群算法的神经网络对企业资信评估有着良好的性能。

关键词: 蚁群算法, 神经网络, 资信评估, BP算法