This PhD project under the supervision will be concerned with new rational Krylov techniques and machine learning techniques in the context of time series modelling. One of the aims is to extend the capabilities of the Rational Krylov Toolbox (http://rktoolbox.org) to estimate and simulate VARMA time series models. Another aspect will be the use of recurrent neural networks for predicting time series based on new symbolic representations. A strong background in numerical linear algebra and programming skills in MATLAB and Python are essential. Good knowledge of statistics is advantageous.
Position is open until filled and applications are considered immediately