Share
المقدرات الحصينة للارتباط الذاتي في انموذج السلاسل الزمنية الموسمية المختلط المضبب == Robust Autocorrelation Estimation of Fuzzy SARIMA Models
Author name:
طاهر ريسان دخيل الخاقاني
Supervisor name:
عبد المجيد حمزة الناصر
General topic:
Administration and Economics
Specific topic:
Economy
Degree:
Doctorate
University:
University of Baghdad - Faculty Of Administration And Economics
Language:
Arabic
University location:
Baghdad
First pages:
07T3805 - p.pdf
Abstract:
غالبا ما تصاحب عملية جمع البيانات وجود قيم شاذة التي تؤثر بشكل كبير ليس فقط على عملية تشخيص الانموذج الملائم الذي يعتمد على معاملات الارتباط الذاتي والارتباط الذاتي الجزئي وانما يتعدى ذلك الى عملية تقدير معلمات الانموذج, ومن هنا فان هذا البحث يهدف الى جعل | The presence of outliers in data influences not only in model diagnostic, but exceed to parameters estimation. Therefore, the goal of this thesis is to make the autocorrelation coefficient robust against outliers to estimate parameters that will be robust depending on those coefficients. This goal is achieved by three suggestions the first and second are depending on Fuzzy Logic by suggest estimation procedure to estimate autocorrelation coefficients called Semi - Fuzzy Autocorrelation coefficient, and suggest another one called Semi - Fuzzy Kendall. The third is to extend the robust Percentage Bend Correlation method to use it in estimate the autocorrelation coefficients that will be used in parameter estimation. This thesis contains four chapters. The first chapter discuss general introduction, the previous studies and the purpose of this research. The second chapter views the general fundamentals, the theoretical field that related with outliers and its effectiveness in time series modeling, the using of fuzzy logic in time series and display the proposed procedure and its algorithm. The third chapter contains the empirical filed, the simulation experiment, results analysis and comparing by some criteria. The fourth chapter views the practical filed by study a daily time series of electrical load in Al - Qadisiya city in Iraq then apply the previous methods. The fifth chapter views the most important conclusions and recommendations that we get.
