Project Outline
EECS 6412
Fall 2019
York University

Projects will be completed in groups of 1-2 students. The topic of a project can be research-oriented or application-oriented. You are welcome to discuss with the instructor about your project topics.

Categories of Projects

  • Research-oriented. Students will define a research problem, develop a reasonable solution and conduct an empirical study on a few datasets.
  • Application-oriened. Students will find a real world dataset, define an application problem, apply at least three exsiting algorithms to solve the problem (or you can design your own solution and compare with at least two existing ones), and conduct an empirical comparison of the algorithms on the dataset.
  • Evaluation-oriented. Students will select a (new) algorithm or system that has not been extensively evaluated. Find or create datasets to evaluate the algorithm or system.

Components of a Project

A project will include the following components:
  • Project proposal. Students should write a project proposal (1 or 2 pages), which describes the project topic, the problem to be solved, the approach(es) taken, the dataset(s) and the project schedule.
  • Design of your method. For a research-orientend project, you may design a new data mining problem and design an algorithm to solve the problem, or you may design a new algorithm to solve an existing problem, or revise an existing algorithm to solve a limitation of the algorihtm. For an application-oriented project, you will design a method for solving the application problem with data mining. You will need to find or create a data set for the application.
  • Implementation and experiments. A group working on a research problem will implement their proposed algorithm, test the program on a number of datasets, and compare the results of the proposed algorithm with the results from at least one existing algorithm (if there exists one). Performce measures used in the comparison depend on the nature of the research problem. A group working on the application track will select at least three existing algorithms to solve the application problem. If you can find the programs for the chosen algorithms, you can use them; otherwise, you will implement them by your own. You will conduct experiments that compare the performance of the algorithm on the application dataset.
  • Written report. Your project report should be written in the format of a decriptive paper. You can use the research papers you have read as models. The report should include the following:
    • Introduction
      • The objective of the project
      • Statement of the problem to be solved
      • Organization of the report
    • Related work
      • previous solutions
    • Approach(es)
      • desription of algorithm(s)
      • illustrate examples
      • theorerical analyses of algorithm(s) if possible
    • Experimental results
      • description of the experiments
      • results
      • discussion of the results
    • Conclusions
      • summarize what was achieved
      • findings from the projects
      • Limitaions and possible externsions
    • Appendices
      • User's manual (if you implement a program or programs)
      • Description of your system design (system components, data structures, control flow structures, how your program handle very large dataset, etc.)
      • Sample input and output.
      • Program limitations and known bugs. (such as the maximum size of dataset your program can take)
  • Class presentation Each group should prepare a 5 to 8-minute oral presentation of your project. (Presentation time: Friday December 20 at 10:00am)

How to hand in

You should use the submit command to turn in your work (including project report, programs that you implement, data sets if not given by me, and anything else that you think you should submit for evaluating your project.) before the due time. If your work is in directory Dir, use

submit 6412 Project Dir

You can also use the EECS Web Submit System to submit your files. Log in with your EECS credentials, select 6412 for course and Project for assignment, and follow the instructions there to submit your files.