Assignment 2 Solution

Instructions

This is an open-ended assignment in which you are expected to write a detailed report documenting your results. Please submit your solution electronically via gradescope, on or before Nov 30 (Tuesday week 10). This assignment is worth 25% of the nal grade.

This assignment may be conducted in groups of 1-4 people. Groups of four are allowed, but possibly consider splitting your project into two groups of two unless you have something in mind in which really bene ts from four members. The marking scheme is the same regardless of your group’s size. Make sure to specify the names of all of your group members when submitting. Submissions should be in the form of a written report, which is expected to be at least four pages (double column, 11pt), or roughly 2.5-3 thousand words, plus gures, tables, and equations. See an example template in the lecture slides to get an idea of the length expected.

Examples of datasets and projects that may be of interest in this assignment will be discussed in the lectures, though you may use any dataset you wish (including the ones we used for Assignment 1). For a selection of datasets that I frequently use, see https://cseweb.ucsd.edu/~jmcauley/datasets.html

Tasks

Assignments will be graded based on their coverage of the following ve components. Examples of what might be included in these sections and previous assignment examples shall be described in more detail in class. Each of the ve sections below will contribute approximately 5 percent of your grade, for a total of 25 percent for the whole assignment.

    1. Identify a dataset to study, and perform an exploratory analysis of the data. Describe the dataset, including its basic statistics and properties, and report any interesting ndings. This exploratory analysis should motivate the design of your model in the following sections. Datasets should be reasonably large (e.g. more than 50,000 samples).

    2. Identify a predictive task that can be studied on this dataset. Describe how you will evaluate your model at this predictive task, what relevant baselines can be used for comparison, and how you will assess the validity of your model’s predictions. It’s ne to use models that were described in class here (i.e., you don’t have to invent anything new (though you may!)), though you should explain and justify which model was appropriate for the task. It’s also important in this section to carefully describe what features you will use and how you had to process the data to obtain them.

    3. Describe your model. Explain and justify your decision to use the model you proposed. How will you optimize it? Did you run into any issues due to scalability, over tting, etc.? What other models did you consider for comparison? What were your unsuccessful attempts along the way? What are the strengths and weaknesses of the di erent models being compared?

    4. Describe literature related to the problem you are studying. If you are using an existing dataset, where did it come from and how was it used? What other similar datasets have been studied in the past and how? What are the state-of-the-art methods currently employed to study this type of data? Are the conclusions from existing work similar to or di erent from your own ndings?

    5. Describe your results and conclusions. How well does your model perform compared to alternatives, and what is the signi cance of the results? Which feature representations worked well and which do not? What is the interpretation of your model’s parameters? Why did the proposed model succeed why others failed (or if it failed, why did it fail)?