论文部分内容阅读
由于遥测降雨系统自身的原因,遥测降雨资料中常有异常值的出现.充分利用降雨分布特征以及抗差统计理论,提出一种三步抗差统计方法探测遥测降雨资料中的异常值.本方法采用Tukey fence统计方法抵御异常值的干扰,用三步的形式以适应降雨资料的分布特征.对面平均降雨量进行的分组,进一步提高方法的探测效率.数据证明,新方法的探测效果较好,且符合水文预报要求.
Due to the reasons of telemetry rainfall system, outliers often occur in the telemetry rainfall data.Taking advantage of the distribution characteristics of rainfall and the statistical theory of resistance difference, a three-step robust statistical method is proposed to detect the outliers in the telemetry rainfall data.This method uses Tukey fence statistical method is used to resist outliers interference, and the three-step form is used to adapt the distribution characteristics of rainfall data. The grouping of surface average rainfall further improves the detection efficiency of the method. The data prove that the new method has good detection effect and In line with hydrological forecasting requirements.