基于改进的BP-LM在中国人口老龄化预测中的应用

由于人口老龄化率的预测具有高度非线性特征，这与BP神经网络能够处理非线性问题的特征相符合，但BP神经网络算法易使解陷入局部极小。基于L-M算法的改进BP神经网络可以有效克服这一问题，而且收敛速度快。通过具体的仿真及实践结果验证了改进BP的有效性，并对未来五年的中国老龄化率进行了预测。

Application of Improved BP-LM Algorithm in Chinese Population-aging Ratio Forecasting

High non-liner is characteristic of population-aging ratio forecasting, and this characteristic is in line with the BP neural network which is capable of dealing with the characteristics of non-linear problems. But neural network algorithm tend to make solution fall into a local minimum. The improved BP-LM algorithm can overcome this shortcoming with fast constrin- gency speed. The paper testified the effectiveness of the improved BP algorithm through a simulation experiment and then forecasted China＇s aging rate for the next five years.