SPH 7460 Final Exam December 18, 2012 page 1 of 4 Three Problems - 4 pages Name: _________________________________________ ===================================================================================== 1. A linear transformation T: R^2 --> R^2 is such that: | 1 | | 1 | | 0 | | 1/2 | T | | = | | , T | | = | |. | 0 | | 1/2 | | 1 | | 1 | a) Sketch the image of the unit square under the transformation T. [6] b) Find the matrix corresponding to T. [8] c) What is the determinant of T ? [4] SPH 7460 Final Exam December 18, 2012 page 2 of 4 Name: _________________________________________ ===================================================================================== 1., Continued d) Find the eigenvalues of T. [8] e) Why is the product of the eigenvalues the same as the determinant of T ? [8] f) Write a SAS program to compute the eigenvalues and eigenvectors of T. [6] SPH 7460 Final Exam December 18, 2012 page 3 of 4 Name: _________________________________________ ===================================================================================== 2. Given a data file with 1000 observations of the random variable X, write a SAS macro which will compute the bootstrap estimate of the 95 % confidence interval for the standard deviation of X. The call to the macro should look like: %boot95(datafile, xvar, xstddev, nboots, cl95, cu95), where datafile is the data file, xvar is the variable X of interest, xstddev is the sample xstddev of xvar, nboots is the number of bootstrap samples, and cl95 and cu95 are the lower and upper 95% confidence limits for the true standard deviation. [35] SPH 7460 Final Exam December 18, 2012 page 4 of 6 Name: _________________________________________ ===================================================================================== 3. Concentrations C in the blood of a toxic substance are a function of a person's body surface area A and the time t after ingesting the toxic substance; specifically, C = (f * A + g * exp(-h * t))^2 + e where f, g, and h are unknown parameters and e is measurement error; e ~ N(0, sigma^2). a) Given a dataset D that has C, A, and t for each of N people, write a SAS PROC NLIN program which will produce the least-squares estimates of f, g, and h. [12] b) Describe how you would test the hypothesis that f = 0. [13]