140.754.01 ADVANCED METHODS IN BIOSTATISTICS IV

Reviews key topics in modern applied statistics. Extends topics of 140.753 to encompass modern semi-parametric and non-parametric methods. Topics include linear, nonlinear and multivariate smoothing, semi-parametric models for clustered data, measurement error models, and statistical learning techniques such as classification, decision trees, and boosting. Emphasis is given both to rigorous methodological development and to practical data analytic strategies. Computational methods designed for semi-parametric inference are presented and relevant software is discussed.

Upon completion of this course, students will be able to: 1) understand modern regression tools such as scatterplot smoothing and additive models; 2) apply additive models to public health studies and compare results with those of standard regression models; 3) extend additive models to account for count outcomes and clustering; 4) understand multivariate smoothing and applications to medical imaging and noise reduction; 5) discuss measurement error models and their application to epidemiological studies; 6) understand statistical and scientific model selection and uncertainty; 7) apply modern statistical learning techniques such as clustering, classification trees and boosting for pattern recognition in complex data sets.