Speaker
Tomás Fernández Bouvier
Description
Our aim is to implement an automated quality control assurance method
for meteorological data. The approach is based on a set of different tests
that probe the spatial, temporal an physical consistency of our data in
a statistical driven method. In order to merge the different outputs, we
propose a statistical framework based on accuracy estimation and opti-
misation of our tests and bayesian multiple evidence merging. Finally we
would set the guidelines for an imputation method of the detected outliers
based on the optimisation of the likelihood yielded by our tests.
| Field of study | Computational Physics |
|---|---|
| Supervisor | Eigil Kaas, Xiaohua Yang and Bjarne Armstrup |
