WPI Worcester Polytechnic Institute

Computer Science Department
------------------------------------------

CS539 Machine Learning - Spring 2007 
Project 6 - Instance-Based Learning and Regression Methods

PROF. CAROLINA RUIZ 

Due Date: Thursday, March 22 2007. Slides are due at 3:00 pm and the written report is due at 4:00 pm. 
------------------------------------------

PROJECT DESCRIPTION

Use Instance-based Learning and Regression techniques to construct classifiers for each of the following problems:

  1. Predicting (1) the class attribute, and (2) a numeric attribute of your choice in the census-income dataset.

  2. Predicting (1) a nominal attribute, and (2) a numeric attribute of your choice in a dataset selected by you. This dataset can consist of data that you use for your own research or work, a dataset taken from a public data repository (e.g., UCI Machine Learning Repository, or from the UCI KDD Archive), or data that you collect from public data sources. THIS DATASET SHOULD BE LARGE IN TERMS OF THE NUMBER OF INSTANCES AND ATTRIBUTES SO IT CANNOT BE ONE OF THOSE INCLUDED IN THE WEKA SYSTEM.

PROJECT ASSIGNMENT

  1. Read Chapter 8 of the textbook about Instance-based Learning in great detail.

  2. Solve Exercise 8.3 of your textbook (page 247). Include your solution in your written report (and not in your oral report).

  3. Read the code of the Instance-based Learning and Regression techiques implemented in the Weka system. Some of those techniques are enumerated below:

    • Instance-based Learning:
      • IB1: nearest neighbor classification
      • IBk: k-nearest neighbors classification. Experiment with several values of k.

    • Other Lazy Learning:
      • LBR: Lazy Bayesian Rules Classifier

    • Regression:
      • Linear Regression
      • LWR: Locally Weighted Regression [In order to run locally weighted linear regression using Weka, use LWL (locally weighted learning) from the Weka's lazy classifiers, and select "Linear Regression" for the LWL's classifier option]

  4. The following are guidelines for the construction of your models:

    • Code: Use the above listed techniques implemented in the Weka system. If you prefer, implement your own code.

    • Objectives of the Learning Experiments: Before you start running experiments, look at the raw data in detail. Figure out 3 to 5 specific, interesting questions about the domain that you want to answer with your Instance-based learning experiments. These questions may be phrased as conjectures that you want to confirm/refute with your experimental results.

    • Training and Testing Instances: You may restrict your experiments to a subset of the instances IF Weka cannot handle your whole dataset. But remember that usually the more training data that you can use, the better. FOR TESTING, you may restrict your test set to 100 data instances or less (not used for training) to reduce the time taken by the experiments using these lazy methods.

  • Preprocessing of the Data: You should apply relevant filters to your dataset as needed before doing the mining and/or using the results of previous mining tasks. For instance, you may decide to remove apparently irrelevant attributes, replace missing values if any, discretize attributes in a different way, etc. Your report should contained a detailed description of the preprocessing of your dataset and justifications of the steps you followed. If Weka does not provide the functionality you need to preprocess your data as you need to obtain useful patterns, preprocess the data yourself either by writing the necessary filters (you can incorporate them in Weka if you wish).

  • Evaluation and Testing: Use with n-fold cross-validation (or with percentage split if the execution time required by cross-validation is too high).

    REPORT AND DUE DATE