Project Requirements
- This project is intended to be a team project, where
students ideally work in pairs. If you are anti-social, I
might agree to let you work individually and in rare cases,
I might agree to a group of 3 or more. However, teams consisting of
one individual will be graded on the same basis as teams
of 2 people, while I would expect significantly more
out of any team of 3 people (and even more for a larger team).
In general, it is likely to be to your advantage to work with one partner.
A written report and class presentation are required for this project.
- The objective is to apply machine learning to an interesting and
challenging problem, ideally in a somewhat novel way.
I'm open as to the subject area, so pick something that you
find interesting.
- If a significant dataset is already available for your project, then I'll
have higher expectations for the machine learning aspects (and especially the novelty).
On the other hand, if you have to do a lot of work just to collect a
useful dataset, then the machine learning aspects can be more routine.
In any case, you will need to conduct extensive experiments with at least
one—and preferably more than one—of the main ML techniques that
was covered in detail in class, and I expect you to also experiment with
at least one technique that was not covered
(or was only briefly covered) in class.
- Project topics must be determined by the
close of business on Thursday, September 29. You need
to get approval from me via email
(mark.stamp@sjsu.edu) by then.
In your email, provide a brief summary of what you plan to do.
A couple of paragraphs will typically suffice.
- On Tuesday, October 4,
each team will give a brief (approximately 5 minute)
presentation of their proposed project topic to the class. Prepare a small number of
slides (like 4 or 5) to present, and be sure that you can display your slides on the
screen without any delay. Your first slide should include the title and team members.
Then you need to discuss what you plan to do (e.g., what algorithms and packages you'll use,
and so on), and you need to tell us where the data will come from, and how you will
deal with the data.
- Your written report is due on TBD,
and we will begin presentations on that day.