Ryan's statistical research interests involve penalized regression modeling approaches, which include well-known methods like the LASSO and elastic net, and also non-convex penalties such as MCP and SCAD and their variants, and structured approaches like the group LASSO. These methods are particularly useful in the analysis of high-dimensional data, where the large number of variables causes most traditional regression methods to be unstable. A major appeal of using regression based approaches, rather than other machine learning tools, for the analysis of high-dimensional data is the interpretability of regression models. Ryan's research focuses on false discovery rate (FDR) based statistical inference for these models, including FDR approaches for penalized linear, logistic, and Cox regression models.
Ryan is also actively involved in various applied research projects, including a project with the Iowa Public Policy Center studying disenrollment from Iowa's recent Medicaid expansion model, and a project data mining Iowa Department of Transportation crash records to learn about risk factors in teen driving.
Education and Degrees
Ph.D, MS - University of Iowa, Department of Biostatistics
BA - Augustana College