Abstract by Brianne Gurney
Signal Processing for Noisy Multi-source Data
With ever improving technology, more and more data are being collected to identify previously unknown relationships from noisy and cluttered data. Particularly, scientists, governments, and communities are interested in using remotely sensed data to identify hidden signals of interest. However, the measured data often contain multiple competing sources not of interest making identification difficult. Additionally, the large amounts of data, while meant to provide additional information, can make computation infeasible. We are interested in identifying a specific but unknown signal from antennae data of three measured channels. Previously used methods do not account for statistical uncertainty and rely on only one measured time series or channel. We seek to apply multivariate statistical methods to better identify the signal of interest in a computationally feasible way.