Probability paper and plotting position of extreme value distribution for distribution selection and parameter estimation

Other Title(s)

رسمة ورقة الاحتمال لتوزيعات القيم المتطرفة لتحديد النموذج و تقدير المعالم

Joint Authors

al-Tahir, al-Saidi al-Mahdi
al-Aswad, Hafiz Abu Bakr Amhamad

Source

مجلة جامعة سبها للعلوم البحتة و التطبيقية

Publisher

Sabha University

Publication Date

2018-03-31

Country of Publication

Libya

No. of Pages

5

Main Subjects

Mathematics

English Abstract

The problem of estimation for tail index in extreme value distributions is very important in many applications.

Statistical methods could be used to select the distribution based on the available data; where the initially specified distributions or family of distributions usually depend on unknown parameters and these parameters need to be estimated.

This paper shows how probability papers plots (PPP) can be used to select the most appropriate distribution among the three types of maximum extreme value distribution (Gumbel, Weibull, Fréchet).

Another objective is to use the PPP for parameter estimation using regression method.

The last objective is to compare between PPP, MLE and PWM methods to estimate the parameters of the selected distribution using two criteria, the mean square error and correlation coefficient.

Results of using daily maximum of temperature in weather data show that the Weibull distribution is the best distribution for each period of this data.

Another result finding is that the PPP provide more efficient estimate for the shape parameter of Weibull distribution than MLE and PWM methods dependent on mean square error.

Data Type

Conference Papers

Record ID

BIM-1208028

• APA
• MLA
• AMA