ECS 189G Term Project

Due date:

Scheduled final exam day (no written final), 11:59 pm. NO LATE SUBMISSIONS.

Problem A

Here you will apply recommender systems methods to a very nonstandard RS problem, analyzing voting data in Congress. Here are the details:

• The dataset is in the UCI Machine Learning Repository, under the title " Congressional Voting Records Data Set."
• Cast this as an RS problem, with the members of Congress being the "moviegoers," the legislative bills being the "movies," and the votes (1 for yes, 0 for non-yes) being the "ratings."
• The goal is to predict the missing votes, coded as '?' in the data.
• Try the various methods from our course.

Problem B

Here you will use another item from UCI, the Drug Review Dataset (Large).

The task here will be different from what we've done. You will not be predicting missing ratings, but instead will be analyzing verbal (i.e. text) ratings.

• Keep in mind that this dataset is physically challenging. It is somewhat larger than what we've been working on (though not "Big Data"), and it is in a more complex format. See this page (material on fread()) for tips.
• Your task here will be to investigate how reliable the verbal reviews are in establishing a numeric rating. Can the given numeric rating be predicted well from the verbal one (plus other variables, such as the disease being treated)? Again, this is just exploring, but we hope to gain insight for other datasets in which there is no numeric rating available.
• As your predictive tools, use at least a linear model and neural networks.
• Your main tool will be sentiment analysis. Various R packages are available, such as RSentiment.

IMPORTANT SUBMISSION DETAILS:

• Your project must be in a file TermProject.pdf.
• You are allowed and encouraged to use the Web or other sources. Of course, all such sources must be cited. Use of human sources is allowed only if they are cited.
• Write as if you are consultant and this is a report to a client. Assume the client is reasonably good with quantitative discussions but does not have backround in RS. You may wish to cite our textbook and/or other sources.
• Include a section (for each problem) discussing the "data wrangling" you needed to do in order to ready the datasets for analysis.
• You must do all your coding in R. This is so that I can easily run your code when I grade your project.
• Your group submits just ONE copy of the report.
• Absolutely NO late reports will be accepted. As you near the deadline, keep submitting what you have so (each one will overwrite the last), so that at least you will get a lot of credit even if you don't finish.
• Submit your report, including all files (.tex, .pdf, R files, any image files, etc.) to my handin site on CSIF (NOT the TA's site), directory 189gproject. The name of your file must be of the form email1.email2....tar , where each emaili is the UCD e-mail address of group member i, e.g. bclinton.gwbush.bobama.dtrump.tar. Note the periods separating fields. Your .tar file must contain only regular files, NO SUBDIRECTORIES!!!! And .tar does NOT mean .tar.gz or .tar.bz2 (or for that matter .rar, which one students used once).
• Double check that you are meeting all the specs.
• NO SUBDIRECTORIES! NO SUBDIRECTORIES! NO SUBDIRECTORIES!
• Make sure that all partners' names are on the report, and that the e-mail addresses in the file name are EXACTLY the official UCD e-mail addresses for the students. These are the addresses at which you have been receiving e-mail from me regarding blog posts. DON'T RISK HAVING A TEAM MEMBER FAILING TO GET CREDIT FOR THE PROJECT -- IT HAS HAPPENED!
• Include a section listing each team member's contribution -- who did what. If a member did not participate, do not include him/ her in this section, and in the .tar file name.
• Include an appendix (\appendix \section{} etc.) for code listings. Also remember to include your code files in the package.
• Did I mention, NO SUBDIRECTORIES!

Grading criteria:

• Groups that put a reasonable amount of time -- and thought! -- almost always receive an A or A+ grade on the project. Groups that do not complete the project usually get a D grade. PLEASE START EARLY!
• As explained in class, groups that do good work on the project receive an extra bonus in their course grades, beyond the weight stated in the course syllabus. It often happens that, say, a B+ grade becomes an A- or even an A.
• Technical content of the work (correctness, thoroughness etc.).
• Adherence to instructions.
• Professional quality of the work: Clear, engaging writing, using correct grammar; it need not (should not) be pretentious, but avoid being too colloquial ("the mean was kinda low"). Presentation need not be fancy, but graphs and tables should be used when helpful.
• A+ grades are very possible, and can have a significant impact on your course grade, letters of recommendation, knighthoods, marriage prospects, coronations, etc.