This year I am experimenting with WebCT to deliver the course.
Course description: In this course students will learn about statistical methods for data that contains more than one variable. Topics include dimension reduction using principal components analysis and factor analysis, case reduction using cluster analysis, supervised classification with neural networks and trees. Inferential methods based on the assumption of multivariate normally distributed population will be discussed. Graphics will be heavily utilized in all topics. Schedule of topics: Introduction Graphics Multivariate Normal Inference for the population mean Handling Missing Values Comparing several population means Principal Component Analysis Supervised Classification Unsupervised Classification Factor Analysis Multidimensional scaling Canonical Correlation Course Goals: The objectives of the course are to help students: 1. grasp the concepts and develop critical thinking in multivariate statistical analysis. 2. learn about multivariate problems. 3. compute analyses using standard statistical software. 4. learn sufficient vocabulary to read further about new methodology. 5. apply the methodology to new problems. Course date: Jan 13, 2002 through May 2, 2002 Location: Howe 1304 Meeting days: MWF 1:10-2:00pm Textbook: Applied Multivariate Statistical Analysis, Johnson and Wichern, 2001 (5th ed) Instructor Information: Dr Dianne Cook firstname.lastname@example.org 325 Snedecor Hall 515 294 8865 Office hours MWF 9:30-10:30am or by appointment Teaching Assistant: Wuyan Zhang, Office hours: Tues 2:00-3:00, Snedecor 305 Computing: Software used will be S-Plus and GGobi on PCs. Information on using each of the packages for particular exercises and sample code will be provided by the instructor. The packages have online help. GGobi is free software for multivariate graphics. S-Plus provides student copies at discount prices but there is also a free version called R. The lab PCs have zip drives. Your own personal zip disk ensures that you have the data available after the lab work for additonal exercises or subsequent labs. You will need to buy zip disk to use in labs for saving data and analysis code for use in subsequent labs. Course Grades: Computed from a total of 200pts made up of bi-weekly homeworks (40pts), small review quizzes, (3@10pts=30pts), one mid-term exam (60pts) and a final exam (70pts). Review quizzes will be randomly given. Homeworks will require some computing.