پديد آورنده :
توكلي، زهرا
عنوان :
تحليل داده هاي طولي نامتعادل از ديدگاه بيزي
مقطع تحصيلي :
كارشناسي ارشد
گرايش تحصيلي :
آمار اقتصادي و اجتماعي
محل تحصيل :
اصفهان: دانشگاه صنعتي اصفهان، دانشكده علوم رياضي
صفحه شمار :
دوازده، [۹۳]ص.: مصور
استاد راهنما :
ريحانه ريخته گران
استاد مشاور :
مريم كلكين نما
واژه نامه :
انگليسي به فارسي; فارسي به انگليسي
توصيفگر ها :
داده هاي طولي نامتعادل , همبستگي پياپي , روش هاي شبيه سازي مونت كارلو زنجير ماركفي , مدل انتقال , مدل اتورگرسيو , داده هاي بازگردنده , مساله درون زايي , مدل سازي توام , گمشدگي
استاد داور :
محمد محمدي، مرجان منصوريان
تاريخ ورود اطلاعات :
1396/11/28
چكيده انگليسي :
Analysis of unbalanced longitudinal data in a Bayesian perspective Zahra Tavakoli z tavakoli@math iut ac ir 2018 Department of Mathematical Sciences Isfahan University of Technology Isfahan 84156 83111 Iran Supervisor Dr Reyhane Rikhtehgaran r rikhtehgaran@cc iut ac ir Advisor Dr Maryam Kelkinnama m kelkinnama@cc iut ac ir 2010 MSC 62J05 62N99 Keywords Unbalanced longitudinal data Serial correlation Marcov chain Monte Carlo methods Transition model Autoregressive model Recurrent events Endogeneity problem Joint modelling Missing data AbstractLongitudinal data includes repeated measurements associated with one or more variables for differentsubjects over time These data are used in various sciences such as econometrics social sciences medicine and agriculture In order to analyze this type of data it is necessary to consider two im portant types of dependency the intra class correlation and the serial correlation The intra classcorrelation is due to the effect of the subject characteristics on the observations of the same subjectand the serial correlation is due to the dependence of observations to their previous observationsover time To analyze longitudinal data various models such as mixed effects auto regressive andtransition models are used In mixed effects models by considering random effects in the structureof the model the intra class correlation is controlled In auto regressive and transition models serialcorrelation among observations over time is considered by assuming auto regressive structure betweentwo consecutive error terms and entering lag responses in the model In many longitudinal studies subjects cannot be present at specified time points to record the responsevalues Thus unbalanced longitudinal data are created To analyze this type of data specifically whenthere is a significant serial correlation among observations it is necessary to consider unequally inter vals among observations in the structure of the underlying model which is dealt in this thesis In some studies times of recording observations are random Recurrent event studies are examples
استاد راهنما :
ريحانه ريخته گران
استاد مشاور :
مريم كلكين نما
استاد داور :
محمد محمدي، مرجان منصوريان
