摘要
城市管理中的违规、违法问题事件一直是困扰城管执法人员的一个痼疾,如何对这些城管问题进行有效预测预警,已成为转变城市管理模式,提高城市管理效率的重要问题。针对城管预测预警问题,本文应用时间序列与BP神经网络相结合的思路,通过指数平滑预测、主成分分析、综合评价警度、神经网络预警等设计了一种基于热线数据的城管问题预测预警理论方法和模型,并对城管问题中比较典型的六大类城管问题发生情况进行预测预警。实例分析表明,模型具有较好的预测预警效果,从而为城管问题预警研究提供了一种新的方法和思路,以期对城管问题预警提供参考。
Violation and illegality of urban management issues in urban management have been serious concern for urban managers for years. How to build an effective prediction and early warning system for urban management issues has been a significant way of changing urban management pattern and improving the efficiency of urban management. Focusing on solving the urban management issues, this paper comes up with a prediction and early warning model based on hotline data following the ideas of combining time series with back-propagation artificial neural networks and also using methods of exponential smoothing, analysis of composition, comprehensive assessment of early warning system and neural network early warning and so on. Also this paper forecasts the occurring situation for six types of common issues in urban management. Practical instances show that this model has good prediction and early warning effect. It provides a new way towards the study of early warning system for urban management issues and brings a referential idea for solving this kind of issues.
出处
《微计算机信息》
2011年第12期1-3,共3页
Control & Automation
基金
项目名称:基于数据挖掘的城管热线数据库预警机制研究
基金颁发部门:北京市城市管理综合行政执法局(Q5011212200902)
关键词
城管问题预警
BP神经网络
指数平滑
ANN模型
early warning for urban management issues
BP Neural Network
exponential smoothing
ANN model
