Date of Award

2-1-2025

Publication Type

Dissertation

Degree Name

Ph.D.

Department

Economics, Mathematics, and Statistics

Keywords

Bayesian estimation; Boxplot-valued data; Interval-valued data; Maximum likelihood estimation; Posterior predictive density estimators

Supervisor

Abdulkadir Hussein

Rights

info:eu-repo/semantics/openAccess

Abstract

This thesis focuses on the inference of two significant types of symbolic data analysis: multivariate interval-valued models and multivariate boxplot-valued data. Chapter 3 explores both frequentist and Bayesian methodologies for multivariate interval-valued data, extending previous univariate and bivariate models. Maximum Likelihood Estimators (MLE) and Bayesian estimators are developed, with their performance evaluated through Monte Carlo simulations and real-world applications to highlight their practical utility. In Chapter 4, we consider inference for data presented in the form of boxplots, a symbolic data type that captures variability within datasets more effectively. We propose novel methods for both parameter estimation and density estimation, with applications demonstrated in climatological data. Specifically, summer temperatures in European countries are analyzed, with comparisons between Bayesian and frequen- tist approaches highlighting the advantages of each method. Both chapters emphasize the potential for symbolic data analysis to manage and interpret complex datasets efficiently. This study contributes valuable methodologies for symbolic data, offering insights for future research and applications in fields such as climatology and beyond.

Download

Share

COinS