Non-Gaussian data assimilation via ensembles: A DA application on tourism demand

Data Assimilation, DA, is the process by which an imperfect numerical forecast is corrected according to real observations. The aim of Data Assimilation is to improve the accuracy of forecast methods estimates, by incorporating observations optimally. The main goal of this research is to develop methods to overcome the limitations of some traditional DA techniques. In particular, the performance of traditional DA methods is greatly a ected in the following circumstances: 1. The prior probability distribution is non-Gaussian. 2. The operator of the observations is non-linear and therefore the probability distribution likelihood is non-Gaussian. The main goals of this research are described below: 1. To develop a Data Assimilation framework wherein prior errors are non-Gaussian. 2. To develop a Data Assimilation framework wherein observational errors are non-Gaussian. 3. To adapt and validate a Data Assimilation scheme for AR-models with data of tourism demand.

• Autores: Beltrán Arrieta, Rolando
• Tipo de recurso: Doctoral thesis
• Fecha de publicación: 2019
• Institución: Universidad del Norte
• Repositorio: Repositorio Uninorte
• Idioma: eng
• OAI Identifier: oai:manglar.uninorte.edu.co:10584/13328
• Acceso en línea: http://hdl.handle.net/10584/13328
• Palabra clave: Datos masivos; Turismo -- Modelos matemáticos
• Rights: openAccess
• License: https://creativecommons.org/licenses/by/4.0/