Andreas’ main research interests fall into two areas of Statistics; time series analysis and distributional approximations. In terms of time series analysis, he currently works in change-point detection problems. The purpose is to improve on current approaches, based mainly on the widely known binary segmentation, that estimate the number and the locations of multiple change-points in data. Consistent estimation of the true change-points in time series is extremely useful when it comes to forecasting future values of the data. Andreas aims on improvements both in terms of accuracy and speed.
With regards to approximations in Statistics, Andreas is interested in finding upper bounds on distributional distances between the distribution of a known quantity and its asymptotic distribution. Such bounds give a quantitative aspect on famous qualitative, approximate results. He has already worked on finding such explicit bounds on the distributional distance between the distribution of the Maximum Likelihood Estimator and its approximate normal distribution. For this purpose, a powerful probabilistic technique called Stein's method has partially been used.
Before joining the LSE as a Postdoctoral Research Officer, Andreas was studying for his PhD degree in the Department of Statistics at the University of Oxford.