Ivo Dinov
UCLA Statistics, Neurology, LONI
, Math/PIC
 Courses SOCR Ivo Dinov's Home SiteMap Software Contact

STAT 233
Spring 2004

Statistical Methods in Biomedical Imaging

Department of Statistics

## Instructor: Ivo Dinov

Project 2
Target Date: Friday, June 11, 2004

Please, submit your project in class or electronically around the target date. See the Project submission rules. On the front page include the following header

• (Project_2)  In this project we should try to employ a different technique for statistical analysis of data. In particular, some people may want to produce R, C++, Java, SPSS or other code/software that implements one of the methods we've discussed in class (e.g., EM, PCA/ICA). Again a sample data set is provided in the following table (you may want to use data from your own lab/PI/research that is more appealing to you or other data which you like/understand better). Try to think of an interesting hypothesis or a set of interesting questions that may be addressed in some form of statistical analysis, different from what you attempted for Project 1. I deliberetely do not give you the detailed background on how/where/why did these data arise from, what we did with it and what we found (there may be expected or unexpected findings). We may work in groups (preferebly mixed, strongly recommended) or on an individual basis. You should try to use/learn some computational tool for statistical analysis. These include, but are not limited to, SOCR, STATA, SAS, SPSS, SYSTAT, R, S-plus, or any other source (for many other examples see SOCR).
• Format: Just as if you were actually writing this as a paper, start with a one paragraph abstract, followed by an intro/background of the problem you have chosen to investigate, methods, results, discussion/conclusion and acknowledgements/references, in that order. Clearly state the problem you have chosen to investigate. List the resources you used to come up with the project and reference all sources you used to complete the project.
• Clearly state your hypotheses, prior to interrogating the data.
• Use some statistical techniques we have discussed, or outside methods to convey whether or not there is statistical evidence in support of your original hypotheses.
• Explicitly state your approach to answer your research hypotheses. Write all formulas/tests/statistics you need.
• Interpret your statistical (numerical) results in a lay back language. Write conclusions and discussions at the end of your report and acknowledge outside help. Describe how this project can be extended in the future.
• I prefer electronic submissions, but if you would rather turn in a paper that would be fine.

 Ivo Dinov's Home http://www.stat.ucla.edu/~dinov Visitor number , since Jan. 01, 2002 © I.D.Dinov 1997-2004