摘要
在对我国中长期汽车保有量预测时,针对传统BP算法的不足,采用遗传算法优化BP算法的连接权值,使优化后的BP网络的训练速度和预测精度得到了有效提高,说明该方法具有较好的实用性和推广价值。
Avoiding the shortcomings of traditional BP algorithm, the authors adopt genetic algorithm to optimize the connection weight of BP network in forecasting the long-term vehicle population in China. The forecasting accuracy and the speed of BP network are improved greatly through (optimization). Thus the combination of genetic algorithm and neural network is conducive to its application and popularity.
出处
《长沙交通学院学报》
2005年第2期79-83,共5页
Journal of Changsha Communications University
关键词
汽车保有量
BP网络
遗传算法
预测
vehicle population
BP network
genetic algorithm
forecasting
