Hydrological sequence non-uniformity diagnosis method based on classification

Abstract

The invention relates to a hydrological sequence non-uniformity diagnosis method based on classification. The diagnosis method includes the steps: acquiring statistical information; performing non-uniformity diagnosis for hydrological time sequence variances; performing non-uniformity for diagnosis hydrological time sequence mean values; judging whether hydrological time sequences are consistent or not. According to the method, hydrological non-uniformity is classified according to the definition of the hydrological non-uniformity and different variety parameters, different statistical methods are selected and used for diagnosing according to the variety parameters, and the method is relatively advanced and avoids repeatability of diagnosis objects and non-uniformity and indeterminacy of diagnosis results caused by diagnosis of a plurality of statistical methods due to fuzziness of inspection objects and inspection purposes. The diagnosis method is universal, applicable to random hydrological time sequences and easy to popularize and apply.

Description

technical field [0001] The invention relates to a method for diagnosing inconsistency of hydrological sequences based on classification, which is a statistical inspection method based on hydrological measured data and a diagnosis method based on inconsistency classification. Background technique [0002] Regarding the diagnosis of hydrological inconsistency, the current technologies mainly include the following types. One is the statistical test of the trend and mutation points of the hydrological time series, and the quantitative evaluation of the hydrological elements of the watershed is carried out according to the test results. There are many diagnostic methods of this kind, and representative test methods include Man-Kendall test, Hurst coefficient, Pettitt test, and Spearman rank correlation method. The second is to integrate a variety of statistical testing methods, and conduct multiple statistical testing of trends and mutation points on a single time series to dete...

AI Technical Summary

Problems solved by technology

Although the non-consistent diagnosis system integrates a variety of statistical methods and tries to improve the reliability of the test results through the results of multiple diagnostic methods, the calculation process is cumbersome, and the reliability of the calculation results is not based on a single test method. Significantly improved

[0004] The disadvantage of the above method is: According to the definition of consistency of time series: the mean and variance of the sequence samples do not change over time, the traditional test method only tests the mean of the sample, and ignores the change of variance

Images