CS 539 Spring 2007

Computer Science Department

CS539 Machine Learning - Spring 2007
Project 7 - Genetic Algorithms

PROF. CAROLINA RUIZ

Due Date: Thursday, April 5th 2007. Slides are due at 3:00 pm and the written report is due at 4:00 pm.

Project Description
Project Assignment
Report Submission and Due Date

PROJECT DESCRIPTION

Use Genetic Algorithms to construct classifiers for the following problem:

Predicting the class attribute in the census-income dataset.
Predicting an 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.

PROJECT ASSIGNMENT

Read Chapter 9 of the textbook on Genetic Algorithms in great detail.

Solve Exercise 9.4 of your textbook (page 270). Include your solution in your written report (and not in your oral report).

The following are guidelines to use genetic algorithms to construct your hypothesis:
- Code: You can use one or more of the following alternatives:
  
  write your own genetic algorithm code implementing selection, crossover, and mutation as they are described in Chapter 9 of the textbook.
  Use the publicly available code at "http://lancet.mit.edu/ga/".
  Use the GA code in WEKA used for attribute selection and extend it as necessary to produce a classifier that includes crossover, mutation, and selection operators as they are described in Chapter 9 of the textbook. See GeneticSearch.java under the attributeSelection.
- 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 GA learning experiments. Try to choose questions that are about the domain and not about GAs or the experimental parameters, but that would particularly benefit from using GAs to answer them. These questions may be phrased as conjectures that you want to confirm/refute with your experimental results.
- Decisions, decisions, decisions: You can choose the strategy and parameters of your algorithm as you prefer. Section 9.3 provides a description of a very nice approach to using genetic algorithms to construct a classifier. I suggest you follow this approach, but you can follow a different one if you wish.
  You should decide how to implement/represent the following notions/parameters in your system. Make sure you explain your design choices clearly in your written report as well as in your oral report (concisely):
  - individuals (i.e., what learning method you will train with GAs and how the hypotheses are encoded),
  - population,
  - fitness (in this part you should elaborate on how individuals are used to predict the target class of test data instances),
  - size of the population,
  - fitness threshold,
  - mutation rate,
  - fraction of population replaced by crossover at each step.
- Training and Testing Instances: The more data instances that you use for training and for testing the better.
- 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 appropriate, 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.

REPORT AND DUE DATE

Written Report.
Your report should contain the following sections with the corresponding discussions:
1. Code Description: Describe the code that you used/wrote. In particular, describe the algorithm implemented by that code. Remember to acknowledge any sources of information/code you used for the implementation of your system.
2. Experiments: For each experiment you ran describe:
  - Which of your 3-5 specific questions/conjectures about the dataset domain you aim to answer/validate with your experiments.
  - Your system parameters: individuals (i.e., hypotheses), initial population, definition of fitness, size of the population, fitness threshold, mutation rate, and fraction of population replaced by crossover at each step, etc.
  - Instances: What data did you use to evaluate the fitness of the individuals in your population?
  - Any pre or post processing done to improve the accuracy of your classifier.
  - Include the resulting hypothesis in your report. Elaborate on what the hypothesis "says".
  - Accuracy of the resulting hypothesis. Remember to explain how you measure accuracy in your system.
3. Summary of Results
  - What was the accuracy of the most accurate hypothesis you obtained?
  - Discuss how this accuracy compares with that of your most accurate results from the previous assignments.
  - Discuss the strengths and the weaknesses of your project.
Oral Report. We will discuss the results from the individual projects during the class on April 5th. Your oral report should summarize the different sections of your written report as described above. Each of you will have 5 minutes to explain your results and to discuss your project in class. Be prepared!
Submission and Due Date.
1. Please submit the following file by email by Thursday April 5th at 3:00 pm.
2. [your-lastname]_proj7_slides.[ext] containing your slides for your oral report. This file should be either a PDF file (ext=pdf) or a PowerPoint file (ext=ppt). Please use only lower case letters in the name file. For instance my file would be named ruiz_proj7_slides.ppt
3. Please bring a hardcopy of your report to class on April 5th, 2007. This written report is due at 4:00 pm that day.

CS539 Machine Learning - Spring 2007 Project 7 - Genetic Algorithms

PROF. CAROLINA RUIZ

PROJECT DESCRIPTION

PROJECT ASSIGNMENT

REPORT AND DUE DATE

CS539 Machine Learning - Spring 2007
Project 7 - Genetic Algorithms