项目名称: 时空序列事件统计过程监测研究
项目编号: No.71301117
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 管理科学
项目作者: 刘亚芬
作者单位: 天津工业大学
项目金额: 20.5万元
中文摘要: 在现代社会中,公共健康与环境卫生事件监测是人们关注的热点问题之一。针对这些事件记录的数据是复杂的时空序列,不仅包含时间信息和空间信息,某些数据还存在明显的空间相关性或时空自相关性。本项目将研究时空序列事件的统计过程监测方法,包括考虑事件的时间特征和空间特征,建立基于事件发生率的非参数统计过程监测方法;建立基于时空扫描统计量的统计过程监测方法,研究时间、空间扫描窗口的动态变化;在此基础上,考虑数据存在的空间相关性或时空自相关性,研究基于时空自回归移动平均(STARMA)模型的时空序列事件统计过程监测方法。这些研究能够充分利用事件监测数据中的时间信息和空间信息,提高监测的有效性和灵敏度。本项目研究成果在理论上是对统计过程控制理论的一个重要补充,在实践上又可应用于公共健康、环境等时空序列数据的监测分析。
中文关键词: 事件监测;统计过程控制;扫描统计量;时空序列;
英文摘要: In modern society, public health and environmental hygiene event surveillance has caused extensive concern. The data of these events is a complex space-time series, including both the temporal and spatial information. Some of the data show obvious spatial correlation or spatial and temporal autocorrelation. This project aims to study the statistical process monitoring methods for event with a space-time series, including non-parametric statistical process control methods for monitoring a event rate in which both the temporal and spatial information are considered, statistical process monitoring methods based on space-time scan statistics of which the the scanning windows vary, and statistical process monitoring methods for event with a space-time series based on the spatio-temporal autoregressive and moving average (STARMA) model in which the spatial and temporal correlation is considered. These studies for event monitoring take full advantage of the spatial and temporal information to make sure the effectiveness of the monitoring and the sensitivity to a quick detection. The study of statistical process monitoring for event with a space-time series is an important complement of the statistical process control theory. The study has demonstrated the potential in the aspect of monitoring the space-time series for
英文关键词: Event Monitoring;Statistical Process Control;Scan Statistics;Space-time Series;