CS 548 KNOWLEDGE DISCOVERY AND DATA MINING - Spring 2012  
Project 4: Clustering

PROF. CAROLINA RUIZ 

DUE DATE: Thursday April 12th, 2012.

• Slides: Submit by email by 10:00 am.
• Written report: Hand in a hardcopy by 1:00 pm.
• Oral Presentation: during class that day.

Project Assignment

• Read Chapter 8 of the textbook in great detail.

• THOROUGHLY READ AND FOLLOW THE PROJECT GUIDELINES. These guidelines contain detailed information about how to structure your project, and how to prepare your written and oral reports.

  Your written report should be at most 10 pages long (including all the graphs, figures, and appendices). The font size should be no smaller than 11pts.

• Data Mining Technique(s): We will run experiments using the following techniques:
  • Pre-processing Techniques: Feature selection, feature creation, dimensionality reduction, noise reduction, attribute discretization, ...

  • Clustering Techniques:
    • Simple K-means
    • HierarchicalClusterer: Make sure to experiment with different "linkType"s
    • EM

• Dataset(s): In this project, we will use the Communities and Crime Unnormalized Data Set available at the UCI Machine Learning Repository. Convert the dataset to the arff format. The arff header is provided in the dataset webpage.

• Performance Metric(s): A major part of this project is to find meaningful ways of evaluating and interpreting the resulting clusters. Devise a variety of approaches to do so, including but not limited to visualization of the resulting clusters (Weka provides very basic visualization), inspection of the clusters' members to find commonalities, use of clustering-specific performance metrics described in the textbook, etc. The deeper your analysis, the better. You might want to extend the Weka code to provide the evaluation/interpretation functionality you need.

• General Comments Focus on experimenting with different ways of preprocessing the data, varying the parameters of the clustering algorithms, and providing in-depth evaluation and interpretion of the results.

• Advanced Topic(s) (30 points): Investigate in more depth (experimentally, theoretically, or both) a topic of your choice that is related to clustering and that is not covered already in this project. This clustering-related topic might be something that was described or mentioned in the textbook or in class, or that comes from your own research, or that is related to your interests, or that appears in a research paper that you find intriguing.

• Project 4 Grading Sheet