Practical machine learning: Project description

Description

The goal of your project is to apply methods we have discussed in class to an interesting dataset and report results. Alternatively, you can explore a method we have discussed in class in more depth. You will be paired with a classmate to complete the project.

Proposal

An initial project description/proposal is due Friday 2/26.

Please describe the project in general terms: describe the dataset and what makes it interesting; what is the goal of analyzing this dataset (predictive performance, interpretation, visualization); what are specific challenges in using this data (large number of predictors/features, large number of observations, difficult to predict task); propose in very general terms what type of methods you might want to use.

If you are working on a specific method, rather than analyzing a dataset, describe the method you will be working on; characteristics of the method you find interesting; what type of questions you want to answer about the method (statistical or computational); what are your goals in working with this method.

Make this description no longer than 1.5 typed pages. Submit by with subject [project proposal].

Project Writeup

In the writeup please include the following:

  1. A description of the dataset and task (you can build on what you wrote for the initial description).
  2. A description of any transformation, filtering, or procedure you performed on the data before analysis
  3. A description of the methods you used, including a discussion of why they are suitable to the task
  4. A report of results, make sure you address the task described in 1 as fully as possible
  5. A discussion of results, and possible future directions for subsequent analysis

Make the writeup no longer than 5 typed pages. Submit by with subject [project report].

The due date for the project is March 29, 2010.