论文部分内容阅读
空间扫描统计量方法是公共卫生监测领域应用非常广泛的空间聚集探测快速算法。其利用传染病监测数据可探测到病例异常增多的局部区域,对可能的传染病暴发做出早期预警。候选聚集区域的预先生成是该方法的一个关键步骤。将现有的候选聚集区域生成方法应用到包含子区域较多的大区域时,可能导致大量候选聚集区域的遗漏,影响探测结果的准确性;或可能生成大量重复的候选聚集区域,导致随后空间扫描计算时间的延长。本文在原有候选聚集区域生成方法的基础上,提出了一种新的快速算法。它以格网点间隔的优化选择,可减少对可能候选聚集区域的遗漏;同时,基于多重排序算法可在较短的时间之内,删除掉原始候选聚集区域集合中的大量重复。通过山东省西南部608个乡镇点的候选聚集区域生成测试,改进的方法可减少候选聚集区域的遗漏,并在较短的时间内删除掉所有的重复候选聚集。
Spatial scanning statistical methods are widely used in the field of public health monitoring space aggregation detection algorithm. The use of infectious disease surveillance data can detect the abnormal increase in the number of local areas, the possible outbreaks of infectious diseases to make early warning. Pre-generation of candidate aggregate areas is a key step of the method. Applying existing methods of generating candidate aggregate regions to large regions with large sub-regions may result in the omission of a large number of candidate aggregate regions, which may affect the accuracy of sounding results or may generate a large number of overlapping candidate aggregate regions, resulting in subsequent space Scan calculation time extension. In this paper, a new fast algorithm is proposed based on the original candidate aggregation area generation method. It optimizes the selection of grid point intervals to reduce the omission of possible candidate aggregated regions. At the same time, based on the multiple sorting algorithm, a large number of duplicates in the original set of candidate aggregated regions can be deleted within a short time. Through the test of the candidate agglomeration area of 608 townships in southwestern Shandong Province, the improved method can reduce the omission of candidate agglomeration areas and delete all the repeated candidate agglomerations in a short time.