DrMarcel Scharth

Marcel's current research focuses on Bayesian methods, Monte Carlo methods, statistical learning, time series, financial econometrics, causal inference, especially the intersection between these areas.

His primary research focus is the development of Bayesian and computational methods for the estimation of complex high-dimensional models, in particular models for dependent data such as time series and longitudinal data.

One of his key topics of research has been the estimation of state space models for time series based on Monte Carlo methods such as high-dimensional importance sampling, sequential Monte Carlo, Markov Chain Monte Carlo (MCMC). Marcel's research has also highlighted the application of these methods in financial econometrics, in particular for the estimation of univariate and multivariate stochastic volatility models.

Marcel's current research also focuses on Bayesian machine learning and statistical learning methods for time series and longitudinal data. This research addresses the challenge of learning complex and possibly time-changing nonlinear regression patterns and lower-dimensional structure from dependent data, while accounting for standard time series patterns, mixed effects, and other features that arise in this type of data.