摘要
以神经网络和时间序列分析方法为基础,采用零均值化、标准偏差预处理方法、规则化能量函数法和贝叶斯规则化方法进行BP神经网络建模,利用BP网络对平整度非平稳时序进行趋势项提取,使非平稳监测时序转化为平稳时序以进行常规ARMA时序分析。结合滚动预测方法,建立了适合平整度预测的神经网络时间序列分析联合模型,并以江苏省某高速公路的平整度数据为例进行了预测分析。研究结果表明:新模型的预测精度高、实时可靠,可应用于实际工程。
Based on the principles of artificial neural network and time series analysis, the BP network is established by zero mean method,standard deviation preprocess,regularization energy function and Bayes-regularization to extract the trend term of displacement time series.After the extraction,the displacement time series becomes a balance series,which could be processed by normal ARMA model.In addition,combined with the real-time tracing algorithm,the artificial neural network-time series analysis（united modeling）for nonlinear displacement in geotechnical engineering is proposed. As a test,this modeling is used in IRI prediction of a highway in Jiangsu Province.The results of engineering case indicate that it is reliable with high precision.It is proved that this modeling can be used in practical engineering.
出处
《现代交通技术》
2007年第3期8-11,共4页
Modern Transportation Technology
关键词
路面
平整度
神经网络
时间序列
预测
pavement
roughness
neural network
time series
prediction
