Tefaruk HAKTANIR

Makale/Derleme | 2022 | Journal of Statistical Theory and Practice16 ( 3 ) , pp.1 - 14

The skewness coefficient (G) of the generalized logistic (GLO) distribution is a function of its shape parameter (a) only. Both the methods of probability-weighted moments and maximum-likelihood (ML) mostly yield magnitudes for the shape parameter much different from that by the method of moments, the gap narrowing with increasing length of the sample series. The computation of ML parameters by the conventional Newton–Raphson method is problematic with no solution for a nonnegligible number of sample series. Here, the three-step Newton–Raphson algorithm, which was previously proposed for the generalized extreme values distribution, . . .is adapted to the GLO distribution, and on many recorded annual flood peaks and annual maximum rainfalls series and through a comprehensive Monte-Carlo experiment it is shown to improve the rate of convergent solutions considerabl Daha fazlası Daha az

Tefaruk HAKTANIR

Makale/Derleme | 2022 | Arabian Journal of Geosciences15 ( 272 ) , pp.1 - 19

The aim of this study was to develop new analytical models to defne the regional rainfall intensity−duration−frequency (IDF) relationship for the Inland Anatolia Region, which is determined discretely by the L-Moments method at many values of rainfall durations and frequencies. First, the parameters of each one of nine commonly used empirical equations were calibrated to provide the best possible defnition of the IDF relationship for the Inland Anatolia Region. Next, analytical models best ftted to the IDF relationship of the L-Moments method were generated by the artifcial bee colony programming (ABCP) approach, such that a combina . . .tion of nine diferent sets were simulated, taking into account three cost functions and three maximum depths. Mean absolute error, root mean square error, mean square error, Nash–Sutclife efciency coefcient, Willmott’s refned index, performance index, and coefcient of determination were computed to assess the accuracies of the empirical equations and of the ABCP models. These criteria revealed that the ABCP models defned the IDF relationship better than the empirical equations over the entire range of frequencies from 2 to 10,000 years. The accuracy of the empirical equations is much worse than the ABCP model, especially for frequencies smaller than 2000 years. Finally, Kruskal−Wallis tests were applied on all of the IDF relationships given by (1) the L-Moments method, (2) the empirical equations, and (3) the ABCP approach. These results indicated that the numerical values of these three models were from the same population Daha fazlası Daha az