Daily Log: Week 8

Test to see if the sentiment analysis works on the UD definitions to guess the sentiment of slang
Testing! Testing!!
Start writing
Update bibliography

Submitted my TAMU application :)
Looking over the results from the tests.
Debugging!!
Rerunning tests with new information:
- Sentiment now returns the total number of words in the message, the total negative score, the total number of negative words, the strength of negativity (total negativity/total number of negative words), and the percent negativity (the total number of negative words/the total number of words in the message)
- Swear words now return +inf as their negative sentiment (never okay to have a swear word)
- Spell checking is definitely slowing things down, but my debugging fixed some errors for sure.
- Note: doesn't handle emoticons.
Try looking at both UD defn and the WordNet synsets (for cases like "blunt")

Finished writing my application. Had Steph take a look at it, and made changes based on her edits.
Fixed up my resume so that it fit on one page!
Wrote a parser that put the HTML kidzworld comments into one cell per message (instead of one line per message)
Re-ran the test on this file
Went to the Summer Scholar's Luncheon. Learned about how to give an effective presentation.
I think line breaks may be messing up the tests in the comments... rewrote parser so that it gets rid of the line breaks and changes them into spaces...
Started analyzing the test results... but then realized the sentiment analysis function was returning a negative number while the other function was returning a positive number (so adding these together could have resulted in a 0 sentiment when in fact it should have been all negative or all positive).
Re-running tests!

Updated my pictures. I went to Austin this weekend! I also sewed a dress with a sewing pattern and a sewing machine. Steph taught me!
Running tests on the kidzworld data
- Ran sentiment analysis on ~38,000 lines
- Changed parser so that all the lines from the same comment are in the same cell in the CSV file. Running sentiment analysis on each comment as opposed to each line (7098 total)
- Multiple x's messages and the "lolololol..." lines were making the program run SUPER slow, so I put in a hack for it. Obviously I need to figure out why it was making the program run so slow though (pretty sure it has to do with the spellchecker)
Outlined my TAMU statement of purpose essay. Ready to write.
Wrote everything but the last few sentences of my application. Eyes are hurting from staring at the comp screen x_X

CSE DREU: Summer 2012