11 Oct 2005
CS 5244 - Computational Document
Analysis
9
Foster’s features
ˇVery
large feature space, look for distinguishing
features:
lTopic
words
lPunctuation
lMisused
common words
lIrregular
spelling and grammar
ˇSome
specific features (most compound):
lAdverbs
ending with “y”: talky
lParenthetical
connectives: … , then, …
lNouns
ending with “mode”, “style”: crisis mode, outdoor-stadium
style