摘要
AFC系统的核心是票/卡售检票自动处理。为保证AFC系统正常运营,需要定期或按需执行票/卡调配。把时序数据分析技术与数据挖掘理论相结合,建立了适合数据挖掘中径向基函数神经网络的输入样本模型。该模型能够通过反复学习从时序数据中发现潜在的规律,并将其用于轨道交通客流量的短期预测。预测结果表明比采用BP神经网络模型的预测结果精度更高、效果更好。
The core of AFC system is the automatic processing of ticket or card issuing and checking.In order to ensure normal operation of the AFC system,there is the need to deploy tickets and cards regularly or on demand.In this paper,by combining time series data analysis technology with data mining theory,a model for input samples of radial basis function neural network is built and which suits the data mining. The model can find potential rules in the time series data through repeatedly studying,and has been put onto the application of forecasting short-term passengers' flow of rail transit.The result forecasted by the model is higher in precision and better in effect compared with being forecasted by BP neural network model.
出处
《计算机应用与软件》
CSCD
2010年第11期160-162,共3页
Computer Applications and Software
关键词
自动售检票系统
票/卡
时序数据挖掘
自回归
径向基函数神经网络
Automatic fare collection(AFC) system
Ticket/card
Time series data mining
Auto-regressive(AR)
Radial basis function neural network(RBFNN)