测绘通报 ›› 2021, Vol. 0 ›› Issue (6): 156-158.doi: 10.13474/j.cnki.11-2246.2021.0196

• 测绘地理信息技术应用案例 • 上一篇    下一篇

时间序列回归预测模型在自动化监测中的应用

刘建军1, 王嘉伟2, 陈琨3, 韩三琪1   

  1. 1. 宁波市轨道交通集团有限公司, 浙江 宁波 310006;
    2. 浙江华展工程研究设计院有限公司, 浙江 宁波 315012;
    3. 湖北省地质环境总站, 湖北 武汉 437000
  • 收稿日期:2020-04-19 修回日期:2020-04-23 发布日期:2021-06-28
  • 通讯作者: 王嘉伟。E-mail:wangjiawei@whu.edu.cn
  • 作者简介:刘建军(1972—),男,高级工程师,主要研究方向为地铁、公路工程技术管理和工程风险管控。E-mail:jianjunliu123@sina.com
  • 基金资助:
    公路地质灾变预警空间信息技术湖南省工程实验室开放基金(kfj50604)

The application of time series regression prediction model in automated monitoring

LIU Jianjun1, WANG Jiawei2, CHEN Kun3, HAN Sanqi1   

  1. 1. Ningbo Rail Transit Group Co., Ltd., Ningbo 310006, China;
    2. Zhejiang Huazhan Institude of Designing and Rerearch Co., Ltd., Ningbo 315012, China;
    3. Geological Environmental Center of Hubei Povince, Wuhan 437000, China
  • Received:2020-04-19 Revised:2020-04-23 Published:2021-06-28

摘要: 在边坡自动化监测过程中,由于传输信号中断、设备故障、电源中断及传感器替换等原因,不可避免地出现监测数据缺失的现象。数据的缺失对后续的边坡稳定性分析及预测带来不确定的因素,使分析结果产生偏差。本文针对边坡自动化监测数据缺失这一现象,采用时间序列回归预测模型对不同数据缺失率的边坡监测数据进行填补,通过填补值与真实值之间的绝对误差与均方根误差判别其在不同数据缺失率的填补效果,得出该模型对缺失率低于10%的缺失数据具有良好的效果,具有一定的实践意义。

关键词: 时间序列, 缺失数据, 填补, 绝对误差, 均方根误差

Abstract: In the process of slope automation monitoring, because of the interruption of transmission signal, equipment failure, power supply interruption and sensor replacement, the phenomenon of missing monitoring data will inevitably occur. The lack of monitoring data brings uncertain factors to the subsequent slope stability analysis and prediction, which makes the analysis results deviate. This paper aims at the phenomenon of missing dynamic monitoring, the time series regression prediction model is used to fill in the monitoring data of slope with different missing data rates. The absolute error and root mean square error between the filling value and the real value are used to judge the filling effect of different missing data rates. It is concluded that the model has good effect on missing data with missing rate less than 10%, and has certain practice.

Key words: time series, missing data, filling, absolute error, root mean square error

中图分类号: