CSCE 990 (Spring 2006) Homework 3

Assigned Thursday, March 30
Due Sunday, April 23 at 11:59:59 p.m.
Total points: 125

When you hand in your results from this homework, you should submit the following, in separate files:

1. A single .tar.gz or .tar.Z file (make sure you use a UNIX-based compression program) called username.tar.gz where username is your username on cse. In this tar file, put:
   • Source code in the language of your choice (in plain text files).
   • A makefile and a README file facilitating compilation and running of your code (include a description of command line options). If I cannot easily re-create your experiments, you might not get full credit.
   • All your data and results (in plain text files).
2. A single .pdf file with your writeup of the results for all the homework problems, including the last problem. Only pdf will be accepted, and you should only submit one pdf file, with the name username.pdf, where username is your username on cse. Include all your plots in this file, as well as a detailed summary of your experimental setup, results, and conclusions. If you have several plots, you might put a few example ones in the main text and defer the rest to an appendix. Remember that the quality of your writeup strongly affects your grade. See the web page on ``Tips on Presenting Technical Material''.

Submit everything by the due date and time using the web-based handin program.

On this homework, you may work with a partner if you wish, and submit a joint report. Note that I will grade your report more rigorously if you do.

1. (120 pts) In the last homework you put your kernel to work, experimenting with various data sets and plugging it into an existing SVM optimizer such as SVM^light with a high level of success. But now your kernel doesn't want to work with that optimizer any longer (artistic differences). Thus it is up to you to build a new one for your kernel.

   Your mission on this homework is to first implement one of the following SVM optimization algorithms from Chapter 10: (i) Chunking from Section 10.4.1, (ii) Working Set Algorithm from Section 10.4.2, or (iii) Sequential Minimal Optimization (SMO) from Section 10.5 (see Platt's original book chapter on SMO for another perspective on the algorithm). Then rerun your experiments from Problem 1 of Homework 2, using your new optimizer in place of the one you used for Homework 2. Everything else is the same: use the same data, same kernel parameters, etc., and report the same things that you did for Homework 2. In addition, compare the results from your optimizer with those from Homework 2, and explain any differences.

   If you implement (i) or (ii), then you'll need to employ a non-trivial selection strategy (not random). I suggest one from Section 10.4.3. Further, for your optimization algorithm, you may either implement your own version of an interior point algorithm or use an existing convex optimization package, such as the one offered by Matlab.

   As before, when you hand in your results, submit source code electronically as well as a well-written, detailed report (much of your grade will depend on your presentation).

2. (5 pts) State the approximate amount of time you spent on each problem.

Last modified 16 August 2011.