8000 Commits · s1n/thesis · GitHub

More Web Proxy on the site http://driver.im/

This repository was archived by the owner on May 19, 2023. It is now read-only.

Commits

Commits on Nov 28, 2010

Commits on Nov 17, 2010

[docs] Few files with figures and images for the final draft.
s1n
committed
[seed] Renamed the stripwords to stopwords.
s1n
committed
[seed] Stopwords are now loaded into a hash structure.
s1n
committed
[docs] Unrelated documents now.
s1n
committed
[conf] Not interested in chasing WordNet 4 levels deep anymore, now that it only counts instances/leaves.
s1n
committed
[results] Updated results to include MSL corrections (mistakenly commented out a line that used the preexisting score). Also includes adverbs from WordNet now.
s1n
committed
[docs] Updated to include a few more sections.
s1n
committed
[corpus] Rename the NASA corpus module to be ASRS, which is correct.
s1n
committed
[corpus] Keep reading until there is nothing left to read. Blank links and junk lines should not be processed.
s1n
committed
[data] Rename the stripwords to stopwords.
s1n
committed
[seed] stripwords is not more appropriately called 'stopwords'.
s1n
committed
[results] Minor corrects for the manual rename.
s1n
committed
[results] Rename the manual annotations to the acronym PCMA: Pipeline Classification Manual Annotations.
s1n
committed
[seed] The weight modification function is increasing too fast; this will slow it down for at least 3 iterations. Comment out the verbose debug statements.
s1n
committed
[seed] Comment out the verbose debug statements.
s1n
committed
[corpus] Sadly, the best way to speed up regular expression searching over a very large corpus is to use egrep and not perl. This should now be able to finish in just over 6 hours.
s1n
committed
[results] Manual annotations from a second human annotator. This data set verifies the original set is accurate.
s1n
committed

Commits on Nov 15, 2010

[docs] More work towards finishing the written thesis.
s1n
committed

0