摘要
针对基于神经网络的财务预警方法网络结构复杂和训练时间长的缺点，笔者提出丁基于粒子群优化神经网络的财务预警方法．首先对样本数据进行归一化处理，然后采用粒子群优化的BP神经网络来进行训练，最后用训练好的神经网络对我国上市公司财务状况进行预测．仿真实验表明，该方法克服了普通BP神经网络的缺点，使得网络结构的复杂度降低，同时提高了预测的准确度．
The financial early - warning based on neural network has shortcomings of complex network structure and the long - time training. This paper presents a BP neural network based on PSO financial early warning methods （PSOBP）. First, the sample data are normalized, and then trained by using PSOBP, and finally by using the trained methods, Chinese listed companies＇ financial position is predicted. Simulation results show that the new method overcomes the common shortcomings of BP neural network, reducing the complexity of network structure and improving the accuracy of prediction.
出处
《山东师范大学学报（自然科学版）》
CAS
2011年第4期19-23,共5页
Journal of Shandong Normal University(Natural Science)
关键词
财务预警
粒子群算法
BP神经网络
financial early warning
particle swarm optimization
BP neural network
