Document Details

Document Type : Project 
Document Title :
Bayesian method for the direct selection of the rank of self-gradient two-dimensional
اسلوب بايزي مباشر لاختيار رتبة عمليات الانحدار الذاتي ثنائية الابعاد
 
Document Language : Arabic 
Abstract : Problems arise for the definition and assessment, diagnosis and prediction of time series when there is a two-dimensional views of the two series linked. And two-dimensional analysis is based on the modeling and study of these two variables together in the same time, in order to understand the nature of the internal relations between them and increase the efficiency and capabilities of statistical prediction. And the definition of the form the first stage and the task when the time series analysis of two-dimensional and is intended to try to choose an appropriate model among the family of linear models for two-dimensional time series. In other words, by definition means choosing the form of appropriate rank among the ranks of possible two-dimensional linear models. And this stage is the most difficult stages and the most dangerous because all the other phases (Assessment - Diagnosis - forecasting) and depends on the efficiency of the model chosen. In general we can say that there is no way that can be called the ideal way to choose the appropriate model. This research aims mainly to develop a Bayesian method to identify and choose the self-rank regression models using two-dimensional distribution of tribal Jeffrey. And the proposed method depends on the employment of function possible to derive the conditional distribution dimensional regression models to the rank of two-dimensional self in an appropriate form can be calculated with a posteriori probabilities for each possible value in a vacuum rank parameter model. In order to verify the performance of the model Albaaza proposed and tested its accuracy in the selection of the rank regression model of self-D has been a study of simulation using three distributions of different tribal, has been shown of the output of this study that the proposed model was able to identify effectively the ranks in the case of regression models of self-two-dimensional The accuracy is high when choosing sample sizes medium and large. 
Publishing Year : 1423 AH
2003 AD 
Sponsor Name : kau 
Sponsorship Year : 1423 AH
2003 AD 
Added Date : Wednesday, April 30, 2008 

Researchers

Researcher Name (Arabic)Researcher Name (English)Researcher TypeDr GradeEmail
سمير مصطفى شعراويsharawi, samir mustafaInvestigatorDoctorate 

Files

File NameTypeDescription
 30482.docx docx 

Download This Page

Back To Researches Page