Assignment 4: Advanced Topics and Project Plans

Start date 7 March, due 10 and 28 March.

1. What are your plans for a final project? Guidelines:
   • It should involve a real system component. Examples would include implementation of a new algorithm, implementation/evaluation of an algorithm from the literature that hasn't yet been implemented/evaluated), or use of data mining to address a problem with real-world data.
   • It should have novelty. For example, association rule mining on market basket data isn't novel (unless you have a new algorithm.) However, if you come up with some new type of knowledge to be gained from the data, or mining of some new data type that nobody has dealt with (successfully) before, a project can have novelty even if you address the challenge by composing/preprocessing/postprocessing and using existing techniques.
   • It should be challenging. You've got six weeks to work on this, and I expect you to be working on it for all six weeks (not full-time, of course.)
   I'd like a one to two page writeup, which discusses:
   1. What the project is (introductory overview.)
   2. Why it meets the above criteria.
   3. What resources you will need to complete the project, and how you will obtain the resources. Include a rough schedule of milestones.
   4. What the final outcome of the project will be (code, paper, ?)
   5. Evaluation criteria. This isn't so much what makes the project a success, but how (based on the outcome) I should evaluate what you've done and your knowledge.

   Group project proposals are welcome. If you are proposing a group project, the evaulation criteria should specify the different roles of the group members, and how your contributions will be evaluated individually.

   I would like a draft sketch (at least the introductory overview mentioned above) by Friday, March 10. I will provide feedback over spring break, meet with you if you desire March 20-22, and then you can provide the final writeup March 28.

2. Advanced topics. For this, you should provide a prioritized list of topics that you would like to cover in class. For each topic, provide at least one source for studying that topic (the most common source would be a paper from the research literature - please provide a link to electronic copy if possible, although just a citation is okay.) I don't guarantee that we'll cover every topic, and I may choose different papers than you suggest for covering a topic, but this should give me a good idea of your interests so that I can do my best to cover what you find most exciting. Note that this doesn't have to be (but may be) related to your project.

   Suggested sources for papers are Data Mining and Knowledge Discovery, Transactions on Knowledge and Data Engineering, Knowledge and Information Systems, SIGKDD Explorations, and the KDD, ICDM, ECML/PKDD, PAKDD, and SIAM Data Mining conferences proceedings. You'll find the latest (and often preliminary) in workshops associated with those conferences, but I'd go with workshop papers only from the last couple of years. You'll also find work in conferences from other areas such as AAAI and SIGMOD, in particular the SIGMOD data mining workshops (probably on a par with some of the conferences above for quality.)

   You might also find the list of topics from last year (below) to be a start.

I would like one or two topics by March 10. The rest can wait until March 28. You should expect to present a paper and lead discussion on one of your topics in class on March 23, 28, or 30 - a preferred date would help.

Turning in assignment

Electronic submission preferred. Please email to . Pdf is the safest for capturing non-text Hard copy is acceptable, please hand in at the beginning of class.