|
Here is the latest semester STA 340-623 Web Site for registered students
This course concerns Bayesian inference and decision theory/analysis. We cover aspects of theory and methods of statistical decision analysis, with problems of inference (estimation and testing) and prediction under uncertainty feeding into decisions, and statistical parameter estimation as decision problems. We explore and develop theory and methodology: probability models for inference and prediction, loss/utility and risk function development, and aspects of personal utility theory and decision analysis foundations. Implementation of statistical models in prediction and decision problems requires numerical/computational methods for model fitting and evaluation, as well as for optimization in decision contexts. The courses covers some relevant theory and methods for this, including the EM algorithm, direct simulation methods, and aspects of constrained convex optimization.
Integrated into the theory and methods are multiple examples in forecasting in business, finance and economics, decision analysis in areas including mixture-model based classification, medical decision making and diagnostic use of mixture models, statistical parameter estimation and testing as decision problems, detailed development of formal decision analysis in financial portfolio contexts, and others.
Students explore concepts, theory and numerical algorithms for more complicated optimization problems, with examples that link to some of the challenges to the application of decision analysis in increasing complicated real-world problems. This involves and relies on major development of Bayesian prior:posterior:predictive analysis in increasingly realistic models, with subsequent evaluation of decision implications, throughout.
Routine homework exercises involve theoretical developments and derivations, model development and computational implementations, simulation and optimization analysis and implementations, and exploration of ranges of applied problems. Midterm exams (2) reflect learning as represented in classes and homeworks. Intensive course projects involve small teams of students (vertically integrated as possible) working on selected areas from the published literatures and/or student group originated projects, with preliminary and then final presentations and written summaries of projects to complete the course.
 |
 |
 |
| Community Standard & Conduct |
Community Standard & Conduct |
Green Classroom Certified |
|
|
|
