CS5014: Homework 6
Due Friday, 6 October
Please turn in a hardcopy of the solution to this homework, written
using LaTeX. You may not receive full credit if you do not staple
together the pages of your homework.
n
Consider the "User Interface Design Study" in course notes page
"Some Uses of Experiments in
Computer Science". Suppose that the user interface for a
particular software application is being studied, and that the
response variable is the time required to complete a set of tasks
using the application.
- Comment on the suitability of each of the following experiment
designs for the study: simple, full factorial, or fractional
factorial.
- How many experiments would be required for a simple design, a
full factorial design, and a fractional factorial design that was a
half-replicate? (For the fractional factorial design that was a
half-replicate, you should assume that the "user experience" factor
has two levels.)
Do problem 17.1 in Jain. Show your work, possibly by attaching
a copy of a spreadsheet, including its formulas. For part (a),
construct a sign table and derive the nonlinear regression model
expressing MIPS rating y in terms of x_A, x_B, and x_C.
If you use a spreadsheet to solve problem 2 on the homework, you
can just attach the spreadsheet with formulas to your solution. If
you work the problem by hand, use LaTeX to generate the table in your
solution.
For each of the four experiments in Table 1
of the WWW4 paper by Abrams, Standridge, Abdulla, Williams, Fox,
("Caching Proxies: Limitations and Potentials"), state the name of the
experiment design that was used (i.e., "Experiment 1 is a 3**4*3 full
factorial design with replications.").