Grace Wahba

I am interested in multivariate function estimation and statistical model building, given large quantities of heterogeneous, noisy observational data, and, sometimes, prior information and physical constraints of various kinds. I am interested in statistical theory, the development of efficient numerical and statistical methods for large and extremely large data sets, and applications in the following areas: (1) biostatistics: identification and flexible quantification of risk factors in large medical and demographic data sets; (2) numerical weather prediction: new methods for ingesting direct and indirect, noisy heterogeneous observational data into global-scale numerical weather prediction models; (3) supervised machine learning: new or improved methods for supervised machine learning with emphasis on complex but interpretable learning models (recent work involves contributions to the development of support vector-reproducing kernel methods for improved flexible classification algorithms); and (4) climate: new methods for analyzing multivariate global historical climate data, building statistical models, and examining data patterns related to global change.