摘要
本文介绍了BP神经网络算法在鸟情时间序列预测中的应用,利用时间序列的自相关系数确定时间序列的变动周期从而确定网络的拓扑结构,进而实现鸟情时间序列的有效预测。实验结果表明该方法在鸟情预测的应用方面较基于统计的方法具有更好的非线性拟合能力,预测精度更好。
This paper describes the application of the BP neural network algorithm in the timeseries prediction of the birds.Through time series autocorrelation coefficient to get the seasonal fluctuation cycle,which determine the topological structure of network,and realize effective forecast of birds timeseries.The result of experiments shows that the method has better nonlinear fitting,better prediction accuracy than the statistics-based approach.
出处
《信息技术与信息化》
2013年第3期93-97,共5页
Information Technology and Informatization
基金
山东省科技发展计划基金资助项目(2011YD01099)
关键词
神经网络
时间序列
预测
Neural network Timeseries Forecasting
