AI Laboratory

The deadline (28.02.99) is over. Please be in touch with Lev Finkelstein to set a date for a meeting.

Attention: the problem is to be able to distinguish ANY mail type, and not only spam mail.

Input data example: spam mail:
        The original mailbox:
               gzip format spam.rawdata.gz, zip format spam_rawdata.zip
        The mailbox in the format of the task (delimiter ---@@@=AILAB=@@@---):
               gzip format spam.ailab.gz, zip format spam_ailab.zip
        The mailbox with headers in the format of the task (delimiter ---@@@=AILAB=@@@---):
               gzip format spam.ailab.with_headres.gz, zip format spam_ailab_with_headers.zip
        Attention:  Browsers on PC get the .gz files from this site with errors.

        Note:   You CAN use the header information if it is present, but your program must also know to handle
                       pure text.
        The converter from mailbox file to the format of the task: spam.c
 
        A batch interface to consult (C4.5): consult2.zip Please read the README file.
 

Questions

Project report template

Helping Material

• Requirements: see the page handed at the first meeting. Get postscript version (gzipped).
• Course Binder can be found in the library.
• MLC++ (Machine Learning Library In C++)
• Lisp Code for ID3 .
• C4.5 classifier (C source and executables, ABSOLUTE path) /net/emc2-31/usr_local/isl/ai/ml/C4.5/
• C4.5 (bibliography)
• Building Classification Models: ID3 and C4.5: some background
• C5.0 (for PC): demo etc
• Introduction to Artificial Neural Networks Multimedia Tutorial
• Neural Networks: Freeware and Shareware Tools
• Shareware and Freeware Neural Network Software
• Machine Learning
• UCI - (Machine Learning Data Bases)
• Lisp - (CMU Artificial Intelligence Repository)