Term Project Guidelines

Each student is required to write a term paper on some subject related to the topics covered in this course. Each student must get his/her term paper subject approved by the instructor.

Abstract Describing Topic

On Thursday, March 20, each student must submit an extended abstract of their research topic and present a short (5 minutes) oral description of their topic to the class. This written summary should be a 2 - 3 page summary of the topic that you have selected. The topic should be very focused. You may address a general topic (for example, genetic programming) but the bulk of your paper will need to address some aspect of this topic in depth (for example, different cross-over strategies for genetic programming). If your paper is strictly a review of relevant literature in an area, then you will be expected to do an in-depth review of the literature in the area (at least 20 references). If your paper includes original work on your part, then you may have fewer references.

You must include at least 5 references in the reference list of your abstract. At least three of these references must come from a refereed journal or referred conference. You must discuss the relevance of these references and cite the references in your extended abstract.

Use the abstract format from Guidelines for Preparing Theses and Dissertations, with the following modifications:

• Change "Date of Degree" to "Date"
• Change "Major Professor" to "Course and Instructor"
• Omit "Pages in Study"
• Omit "Candidate …"

Term Paper Contents

The term paper should consist of:

• title page (name of term paper, name of student, name of course, date submitted)
• abstract (one-half to three-quarters page overview of paper)
• body of paper (pages must begin to be numbered here; must consist of 10-20 type-written, double-spaced pages, 1 inch margins at side, 1.25 inch top margin, 1 inch bottom margin.)
  The arrangement of the body of the paper may differ from paper to paper, depending on the subject and the student's choice in how to logically arrange the paper. Typically, however, it should begin with an introduction (1-2 pages of background information) and conclude with a summary. The summary is a very important part of the paper, for this gives the student the opportunity to draw conclusions, point out patterns that he/she has perceived during the research for the paper, etc. A strong summary can greatly influence the student's grade; so can a weak summary!
• list of references
  The list of references should be in IEEE style as described above. Be sure that you cite all your references in your paper. A reader should be able to easily distinguish between those ideas and thoughts that are your own and those that you have extracted from some reference. If you are exactly quoting from a source, then be sure to use quotation marks. The style for citations within the body of the paper should follow the IEEE style. All references cited in your paper must be listed at the end of the paper.

You are responsible for ensuring that your spelling, punctuation, and sentence structures are in good shape before submitting your paper. Poor English will result in a lower grade.

The term papers are due at the beginning of class on Thursday, April 24. Term papers will not be accepted late.

Term project ideas

Each of the paper ideas below gives a very broad topic that would need to be narrowed substantially to be a suitable term project.

• You may explore the application of machine learning to a new problem of interest to you. It will be important that you carefully describe the source of your data, why the problem is important, what machine learning approaches you considered applying to the problem, and describe and analyze your results.
• You may conduct an in-depth comparison of several learning algorithms or variations on a particular learning algorithm on several data sets. For example, you might want to compare the effectiveness of different methods for dealing with missing attributes in decisions trees or compare different data normalization strategies when using a back-propagation neural network. For this type of research, you will need to review literature relevant to the subject of your comparison, design your experiments carefully, completely describe and justify the experiments in your paper, and present and analyze your results.
• Fuzzy neuro-systems
• Artificial immune systems
• Scalability issues related to machine learning-you may want to investigate methods that could be used to parallelize some machine learning algorithm and report results dealing with speedup and accuracy.
• Ensembles applied to unsupervised learning
• Spiking neurons
• Biological basis of machine learning algorithms.
• Genetic algorithms applied to fuzzy data mining
• Machine Learning Projects from Brown
• Machine Learning Projects from the University of Ottawa
• Student Projects at the University of Bristol
• Machine Learning Courses