摘要
探讨了位置数据异常分析的一种新思路,给出了面向异常的时空表达模型,建立了基于区域位置点统计的异常检测和分析框架,探寻异常在事件影响下的时空发展趋势和分布特征。通过纽约两个月的出租车出行数据的实验分析验证了框架的有效性,并发现一些有趣的规律,为公共安全、交通管控应急、城市功能区分析等提供指导信息。
In this paper, a new method to analyze outliers in the position data was discussed, an anomaly-oriented spatio-temporal expression model was put forward, and the anomaly detection and analysis framework was estab- lished based on regional point location statistics which was used to explore the spatio-temporal distribution charac- teristics of anomaly under the influence of events. Through the experimental analysis of the taxi travel data in New York during two months, the validity of the framework was verified, and some interesting patterns were found to provide guidance information for public security, traffic control on emergency and urban functional area analysis.
出处
《测绘科学技术学报》
CSCD
北大核心
2017年第3期320-324,330,共6页
Journal of Geomatics Science and Technology
基金
国家自然科学基金项目(41571394)
关键词
时空数据
异常检测
事件
空间统计
趋势分析
spatio-temporal data
anomaly detection
events
spatial statistics
trend analysis