CS 3194 Design Project

Due on or before April 30, 2003, 4:30 p.m.

This problem is to be done on a group basis following the Trinity University Academic Integrity Policy. Laboratory problems should be submitted electronically (e-mail to cs3194@ariel.cs.trinity.edu) on or before the due date and should contain a problem write-up, source code to any programs and data sets used in solving the problem. The submitted files should be ASCII text files having Unix end-of-line characters (please convert all Windows and Mac text files to Unix format-I have found that Emacs or the Stone Text Tool seem to do a reasonable job of such conversions). If several files need to be submitted, put them in a directory having name your-last-name-problem-set-number and create a tar archive of this file system and attach it to your e-mail problem submission.

Introduction

The third year design project is to design a system for filtering e-mail messages. By filtering, we mean analyze e-mail headers and/or content and determine what classification should be assigned each e-mail message.

Design teams must:

Research the e-mail classification literature.
You should study existing designs and implementations for e-mail classification to determine their features and limitations.
Decide what a better design means in the context of their research effort.
Complete a prototype implementation of an e-mail classification system which contains the features which make a better system. In this context, the terminology prototype must be a functional prototype. User interface considerations are considered secondary in this problem.

Design Groups

There are 31 third year design students. The following J expressions were used to select 5 design groups.

   names =: 'Arnold, Katherine M.';'Austin, Ryan E.';'Baker, Chris B.';'Barth, David A.';'Caille, Ryan P.'
   names =: names,'Calcaterra, Gina M.';'Choo, Tai-San';'Crowe, Loren D.';'DeFelice, David C.';'Hohimer, Christopher J.'
   names =: names,'Houck, David P.';'Leeds, Gregory O.';'Martin, Jess A.';'Martin, Victor'
   names =: names,'McKenna, Ryan';'Michalson, Brett P.';'Miller, Travis A.';'Nasr, Carmen T.'
   names =: names,'Nevin, Daniel R.';'Papazian, John H.';'Quilici, Bryan D.';'Raisani, Jabbar K.'
   names =: names,'Riela, Sandra J.';'Rossmeisl, Joshua S.';'Rubio III, Robert';'Salter, Eric L.'
   names =: names,'Scholz, Ryan B.';'Schwartz, Scott L.';'Schwarz, Paul D.';'Smith, Christopher A.';'Stoeck, Kieth R.'
   names =: names,'Wack, Rosellen K.';'Wilbraham, Ryan M.';'Wilmes, Theodore R.'
   groups =: 'Group 1';'Group 2';'Group 3';'Group 4';'Group 5';'Group6'
   (/: /: (28$1 0 0 0 0 0 0),12$1 0 0 0 0 0) { ((34?34){>names),>groups
Group 1                
Barth, David A.        
Smith, Christopher A.  (Group Leader)
Scholz, Ryan B.        
Schwarz, Paul D.       
Papazian, John H.      
Raisani, Jabbar K.     
Group 2                
Rubio III, Robert      
Calcaterra, Gina M.    
McKenna, Ryan          
Miller, Travis A.      
Martin, Victor         
Caille, Ryan P.        
Group 3                
Austin, Ryan E.        
Crowe, Loren D.        
Michalson, Brett P.    
Choo, Tai-San          
Stoeck, Kieth R.       
Martin, Jess A.        
Group 4                
Baker, Chris B.        
Quilici, Bryan D.      
Nevin, Daniel R.    (Group Leader)
Wack, Rosellen K.      
Wilbraham, Ryan M.     
Leeds, Gregory O.      
Group 5                
Schwartz, Scott L.  (Group Leader)
Arnold, Katherine M.   
Houck, David P.        
Wilmes, Theodore R.    
Nasr, Carmen T.        
Group6                 
Hohimer, Christopher J.  (Group Leader)
Riela, Sandra J.       
DeFelice, David C.     
Salter, Eric L.        
Rossmeisl, Joshua S.

Each group must meet and select a group leader before the beginning of the second class period (January 22, 2003) and e-mail the instructor cs3194@ariel.cs.trinity.edu the results of this selection.

Design groups must submit preliminary design documents to the instructor by e-mail cs3194@ariel.cs.trinity.edu before the beginning of the fourth class period (February 12, 2003).

E-Mail Classification Designs

The idea of a better e-mail classification system must be independently formulated by each design group. It is expected that groups will engage in a competition with each other in much the same fashion as companies might if they were designing and marketing competing e-mail classification software products.

Design Project Solution [ HTML ] [ PS ] [ PDF ]

2003-03-05