Adaptive Filtering and Nonlinear Models for Post-processing of Weather Forecasts
Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Kalman filters have been used by SMHI to improve the quality of their forecasts. Until now they have used a linear underlying model to do this. In this thesis it is investigated whether the performance can be improved by the use of nonlinear models such as polynomials and neural networks. The results suggest that an improvement is hard to achieve by this approach and that it is likely not worth the effort to implement a nonlinear model.
Place, publisher, year, edition, pages
2015. , 40 p.
IdentifiersURN: urn:nbn:se:liu:diva-119642ISRN: LiTH-ISY-EX--15/4876--SEOAI: oai:DiVA.org:liu-119642DiVA: diva2:825407
Subject / course