Statistical solutions for and from signal processing

With the wide range of fields engaging in signal processing research, many methods do not receive adequate dissemination across disciplines due to differences in jargon, notation, and level of rigor. In this thesis, I attempt to bridge this gap by applying two statistical techniques originating in signal processing to fields for which they were not originally intended. Firstly, I employ particle filters, a tool used for state estimation in the physics signal processing world, for the task of prior sensitivity analysis and cross validation in Bayesian statistics. Secondly, I demonstrate the application of support vector forecasters, a tool used for forecasting in the machine learning signal processing world, to the field of structural health monitoring.Science, Faculty ofStatistics, Department ofGraduat