Assignment 1 (1%)[pdf]
Assignment 2 (10%)Due on Friday September 30th. If you have questions let me know asap.
Assignment 3 (10%)Due on Friday, November 11th. If you have questions let me know asap.
SOLUTIONS to Q2:
Test 1 (20%)Friday, October 7th, 50 minutes.
- Here is an old midterm. We can take part of Wednesday's class to discuss this test.
- The coverage of our test will be everything that we've done upto and
including last Friday's lecture. In particular
- three different types of confidence sets (ID and derive)
- EDF: theory, practice, confidence regions
- simulation of coverage probabilities
- statistical functionals (derive asymptotic confidence regions for linear and "nearly linear" statistical functionals, describe variance estimators, state "sans proof" limit theory for the quantiles)
- parametric and nonparametric bootstrap to estimate simple quantity
(eg. variance) of estimator
Test 2 (19%)Friday, November 25th, 50 minutes. The test will consist of three essay-style questions, although it is not necessary for you to answer in essay form. The grading, however, will be more "essay-styled" as it will depend on the quality of the answer. The questions will be
- Describe nonparametric regression.
- Describe nonparametric density estimation.
- Does the bootstrap always work?
FINAL PROJECT (40%)
- Data summary due on October 31st.
- Plan of project due on November 14th. [information]
- Final project due on December 14th by 5pm (in assignment box), or earlier.
Potential sources of data:
- Some of the LW textbook data would be OK for you to work with.
- Microarray gene expression data for yeast with asssociated paper.
- CISL Research Data Archive
- Data from Peter Guttorp's book. You'll need the book to find out what the data sets are, and if any of these are of any use.
- Statistics Canada
- More Stats Can.
- The World Bank
- google search results
- General life tables (could be Canada, could be elsewhere, may be interesting).
- Dementia (don't know if the site has data though...)
- I don't own the rights to the next two data sets, so I will only post pseudo-versions of these data:
- Cytotoxicity data from this paper (Table 1).