Homework Submission Guidelines
The submission for hw3 should include the following files:
1. your script for deriving unigram and bigram statistics
2. the top 20 probability unigrams and the top 20
bigrams, both for your own corpus and the class corpus.
3. your readme should contain full explanations
of what you did for problem 1. "See the script explanations"
will be worth 0 points.
4. as in 3., include discussion of the features
you decided to use
5. Submit the scripts for extracting the chosen features
from the corpus. Submit the files with ripper features + the Ripper .hyp
file
6. In the readme file, include a short analysis of the results.
- Create a plain text README file (not a .doc file!), including the
following information: Your name, a list and description of the files you
are submitting, a clear description of how to run your program, a description
of your program's design and any additional information that you consider
significant.
- Create a submit directory.
- Copy the source code files into the submit directory and include
all the other files that are necessary for your program to run.
- cd into the submit directory.
- Mail your files to ani@cs.columbia.edu using the command:
$ tar cvf - . | compress | uuencode
temp_file | Mail -s "submit cs4705 hw3" ani@cs.columbia.edu
After a short time you will get an automatic acknowledgement of your submission.
Please note:
- If you do not get an answer after a few minutes, then your program
did not go through. Please resubmit
- If you do not get back a listing of ALL your files (please check
the file sizes to ensure that everything arrived without any problems) then
resubmit!
If you submit once, and then decide to submit again, then your second submission
overwrites the first. All the files from your first submission will automatically
be wiped out.
Late Homework Policy
No late submissions will be accepted to HW3.