摘要
引用统计学中有关统计检验的概念,分析作为插补资料的相关关系条件,并尽可能给出判别准则及判别式。在相关插补水文序列时,首先需要进行以下检验:(1)残差检验;(2)线性相关程度检验;(3)展延范围的确定。残差检验可论证随机误差是否服从正态分布及是否相互独立。线性相关程度检验可分析所假定的线性回归是否合理及相关程度是否满足插补要求。通过插补范围的估计可求出最大的外延幅度,使插补值的取值范围或误差控制在一定的精度之内。最后利用嘉陵江支流渠江荀渡口、罗渡溪两站的实测资料进行检验分析,证明本文提出的计算公式具有较好的精度。
Introducing the concept of statistic test:The correlative requirements forthe hydrologic data extrapolation were studied, resulting in its criteria and discriminats. Incorrelatively extrapolating hydrologic data,the following statistic tests must first be made:(a) residual error test, (b) linear correlation level test, (c) determination of extrapolatingrange. The random errors' characteristics in normal distribution and dependence may be examined by residual error testing. Both reasonableness of assumed linear regression equationsand meeting of correlation level to extrapolation requirements may be analysed by means oflinear correlation level test. From the estimation of extrapolating range,the maximum extrapolation limit may be obtained,thus,the extrapolating errors could be controlled within an allowance. Observed data of both the Xundukou and the Luoduxi hydrologic stations located onQujiang River,a tributary of the Jialing River,were examined by the calculating formula presented in this paper. It is considered that this formula is of good accuracy.
出处
《人民长江》
北大核心
1995年第4期15-19,共5页
Yangtze River
关键词
水文系列
相关分析
插补条件
检验
Interpolation and extending of hydrological data Correlation Correlation analysis Inspection