基于聚类归因的水文数据筛分方法研究

Research on Hydrological Data Screening Method Based on Cluster Attribution

摘要:

为提升水文规律理论研究的效率,有必要对数量多、维度广的长系列水文数据进行筛分处理,以更好地适配研究重点和实际需求。文章以派口水文站为例,基于多种水文特征指标,利用K-均值聚类与归因分析的水文数据筛分方法,对输沙率精测数据进行筛分。结果表明:基于聚类归因的水文数据筛分方法可充分考虑各类水文要素及特征指标之间的相关关系,使水文数据筛分成果更加符合自然科学原理、更具实际应用价值。

Abstract:

In order to improve the efficiency of theoretical research on hydrological laws, it is necessary to screen long-series hydrological data with large quantity and wide dimensions, so as to better adapt to research priorities and practical needs. Taking Luokou Hydrological Station as an example, this paper uses the hydrological data screening method combining K-means clustering and attribution analysis to screen the accurate measurement data of sediment transport rate based on a variety of hydrological characteristic indicators. The results show that the hydrological data screening method based on cluster attribution can fully consider the correlation between various hydrological elements and characteristic indicators, making the hydrological data screening results more in line with the principles of natural science and more practically applicable.