Aristotle University Of Thessaloniki (AUTH)
Στοχαστική μοντελοποίηση και δυναμική εκτίμηση παραμέτρων με εφαρμογές στην επιδημιολογία
Abstract
dc:descriptionThe modelling of epidemic phenomena has a central role in the field of epidemiology, with many new methods providing reliable estimates of disease spread. Deterministic approaches usually fail to fully describe the evolution of epidemics, due to the inherent uncertainty that characterizes the dynamics of these phenomena. Therefore, the transition from a deterministic to a stochastic perspective is necessary. In Chapter 1 an introduction to mathematical modelling in epidemiology is presented, covering the fundamental deterministic and stochastic models, as well as the basic properties of Bayesian filters and Markov chains. In Chapter 2 the innovative aspects of each chapter are highlighted. In Chapter 3 the stochastic SIRD (Susceptible (S), Infectious (I), Recovered (R), Deceased (D)) model with imperfect immunity is presented, based on a continuous-time Markov process. Formulas and algorithms are developed for calculating probabilities and moments of time-related quantities, such as the disease extinction time, alarm time, mortality and infection time of an individual. Through sensitivity analysis, the behavior of these characteristics is examined with respect to the epidemiological parameters of the model. In Chapter 4 a Markovian SIHRD model (Susceptible (S), Infectious (I), Hospitalized (H), Recovered (R), Deceased (D)) is established, based on a four-dimensional Markov chain, focusing on the number of hospitalizations during the evolution of an epidemic. Unlike Chapter 3, emphasis is placed on size-related stochastic characteristics, such as the total number of hospitalizations, the maximum number of hospitalized individuals, as well as the joint distribution of hospitalizations and infections. In Chapter 5 the aim is to combine formulas for calculating the distributions and moments of stochastic characteristics, derived via the Markovian approach, with a dynamic parameter estimation method such as particle filtering. The Markov chain is adapted to the structure of the compartmental SPIR model (Susceptible (S), Presymptomatic (P), Infectious (I), Recovered-Deceased (R)). Stochastic features are examined, such as the number of infectious individuals until the epidemic’s extinction, the time of occurrence of a specific number of deaths, and the number of infections caused by a presymptomatic or infectious individual. The aim is to provide a more accurate description of the time-varying behavior of epidemics. The proposed method was examined on monkeypox data from Ghana, showing improved estimates compared to the methodology with fixed parameters. Finally, by conducting sensitivity analysis and comparing the behavior of the aforementioned characteristics, conclusions are drawn regarding the impact of infectious and presymptomatic individuals on disease transmission.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Papageorgiou, Vasileios
- Παπαγεωργίου, Βασίλειος
Subjects
dc:subject × 30- Στοχαστική μοντελοποίηση
- Εκτίμηση παραμέτρων
- Μαρκοβιανά μοντέλα
- Μπεϋζιανές μέθοδοι
- Φίλτρο σωματιδίων
- Επιδημιολογία
- Κατανομές
- Stochastic modelling
- Parameter estimation
- Markov chains
- Bayesian methods
- Epidemiologic analysis
- Distribution
- Particle filters
- Μodélisation stochastique
- Estimation des paramètres
- Chaînes de Markov
- Μéthodes bayésiennes
- Analyse épidémiologique
- Filtres particulaires
- Μαθηματικά
- Φυσικές Επιστήμες
- Στατιστική και Πιθανότητες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Βιοπληροφορική
- Mathematics
- Natural Sciences
- Statistics and Probability
- Computer and Information Sciences
- Bioinformatics
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/60635