\n
![](\"http:\/\/journal2.alanv.org\/images\/pipe.jpg\")
<\/p>\nYeah so, today I got into work, wrote the final code for the number extraction, wrote a test suite for the entire parser, and hit run.
\n…
\nAnd got a ClassCastException.
\nA little bit of WTFing, and I discovered that what I thought were serialized Classifiers were actually serialized Results. Gaah. Spent about an hour regenerating the actual classifiers (luckily I got a new, faster computer so the experiments didn’t take 2 days…).
\nThen I hit run again.
\nJust like that, I had an 85% accuracy on the solution classification. I can raise that by improving my preprocessing (the way I’m currently tagging things is faulty). That’s on my todo list for tomorrow. Then, it’s tracking down the errors and seeing if they are easily fixable.
\nNotes to self: Add in negative ability (- sign, etc) and allow for multiple numbers in a single span.<\/p>\n
So yeah… in the 5 hours I had at work today, I accomplished more than I had for the past 6 work days.<\/p>\n
Came home feeling completely happy with myself. Amazing how much work manages to affect my mood, even when I don’t think it does. Yesterday I was feeling kinda meh because I hadn’t gotten much (read: anything) done at work. Today I got a lot done. Yay!
\nCan’t help but feel like I’m still a little behind, but whatever.<\/p>\n
Pounded out a prelimary card list for Student Wars: Internships… not that anyone really cares. Will have cards for the PSLC interns and introduce the PSLC faction… which will do something that I’m not quite sure of yet :-P<\/p>\n
Pipe! Whee!
\n<\/p>\n
Oh, and PICKLES!!!<\/a> AHHHHH!! And thanks to chrisamaphone… WTF WTF WTF<\/a>…<\/p>\n","protected":false},"excerpt":{"rendered":" Yeah so, today I got into work, wrote the final code for the number extraction, wrote a test suite for the entire parser, and hit run. … And got a ClassCastException. A little bit of WTFing, and I discovered that … Continue reading
\nXD XD XD<\/p>\n