Temple University
CIS 595 - Graduate Seminar
Knowledge Discovery and Data Mining
Fall 2000

Guidelines for presentations

1. Specifications:

Each presentation will cover 2 related papers
Duration: 40 min, + 10 min questions for each paper
It is acceptable to present papers from the list of papers for your project
Foils may be hand-written - drawings are *highly* recommended. Contact the instructor if you want latex slide templates; computer/powerpoint presentations are also welcome

2. Suggested steps:

Obtain the approval of the instructor about the topic and the papers
Fully understand the papers. You may want to check a few of the references too
Contact the instructor for questions or for a brief 'dry run'
Give a draft of the foils to the instructor(~ 1 week ahead)

3. Typical structure of a presentation

Number in brackets indicate suggested number of foils:

what is the problem [1]
motivation: why is it important / applications [1]
method [3-5]

main idea / intuition
brief sketch of algorithms
performance results

repeat for other methods if exist [3-5]
how do the methods compare to each other [1-2]
weak points of each method [1]
your suggestions for improvements [1-2]
conclusions [1]

4. Grading scheme:

10% clarity of problem description
10% strength of 'motivation'
50% clarity of description of each method
20% comparison of methods
10% suggestions for improvements