Open Access. Powered by Scholars. Published by Universities.®

Physical Sciences and Mathematics Commons

Open Access. Powered by Scholars. Published by Universities.®

Statistics and Probability

Graduate Research Theses & Dissertations

Theses/Dissertations

2019

Bayesian

Articles 1 - 2 of 2

Full-Text Articles in Physical Sciences and Mathematics

Exploring A Bayesian Analysis Of Opinion Dynamics Using The Approximate Bayesian Computation Method, Jessica L. Bishop Jan 2019

Exploring A Bayesian Analysis Of Opinion Dynamics Using The Approximate Bayesian Computation Method, Jessica L. Bishop

Graduate Research Theses & Dissertations

Social media has created a whole new framework in the way we understand ones expression of opinion, and how ones' opinion can influence others. Models of opinion dynamics, such as a probabilistic modeling framework of opinion dynamics over time are given by Abir De, Isabel Valera, Niloy Ganguly, Sourangshu Bhattacharya, and Manuel Gomez Rodriguez in ``Learning and Forecasting Opinion Dynamics in Social Networks." In this paper, we will continue to explore their models, now coming from a Bayesian statistical standpoint, specifically looking at the Approximate Bayesian Computation (ABC) method for the computation of better estimations for the data. We will …


Evaluation Of Epilepsy Surgery Using Bayesian Multinomial Regression, Kacy Danielle Kane Jan 2019

Evaluation Of Epilepsy Surgery Using Bayesian Multinomial Regression, Kacy Danielle Kane

Graduate Research Theses & Dissertations

We are exploring the effectiveness of brain surgeries that are supposed to eliminate or reduce the frequency of seizures in young Epilepsy patients. The long-term effectiveness of brain surgeries is evaluated by ordinal categories and brings longitudinal categorical responses.

Using a Bayesian multinomial regression model we examine the responses by the lobe of brain and other covariates as well as time. To overcome computational difficulties we utilize latent variables for multinomial responses and compare the results with frequentists methods.