企业破产预测的 BP-th 神经网络系统

BP th Neural Network System on Prediction of Business Bankruptcy

论述以双曲正切函数取代ＢＰ神经网络的Ｓｉｇｍｏｉｄ作用函数的效果，验证表明：提高了网络运行效率及输出信息精度．在此基础上建立了企业破产预测ＢＰ－ｔｈ神经网络系统，阐述如何建立企业财务信息的神经网络知识表达模式，通过考核，系统输出的企业破产与否的结论正确．

In this paper sigmoid function of BP neural network is substituted by Th function. The proceeding effciency of network and outputing in formation precision are confirmed. On this basis, BP th Neural network of business bankruptcy prediction is established. The creation of expressingmodel of business bankruptcy prediction on neural network is indroduced. The output dawing from the system on whether or not the business bankrupts is right.