C02-1130 |
adapting the hierarchical decision
|
list algorithm
|
from ( Yarowsky , 2000 ) to our
|
W02-1027 |
several experiments with a decision
|
list algorithm
|
to explore the usefulness of
|
W02-1027 |
Discussion We have tested the decision
|
list algorithm
|
on our annotated corpus , employing
|
S01-1028 |
set suitable for the Decision
|
List algorithm
|
. We detected some reasons that
|
S01-1028 |
will briefly explain the Decision
|
List algorithm
|
in Section 2 . Section 3 will
|
W99-0613 |
characteristics of the decision
|
list algorithm
|
presented in this paper . ( Riloff
|
W02-1003 |
comparing the various decision
|
list algorithms
|
, we also tried several other
|
P95-1026 |
illustrated in Figure 3 ) . The decision
|
list algorithm
|
resolves any conflicts by using
|
C02-1112 |
feature sets as used by the decision
|
list algorithm
|
. Instantiated Grammatical Relations
|
W02-1003 |
more accurate than the decision
|
list algorithms
|
. Many of the problems that probabilistic
|
P94-1013 |
whether the corpus or decision
|
list algorithm
|
was correct in two cases of disagreement
|
P06-2022 |
motivating his use of the decision
|
list algorithm
|
. In contrast , the goal here
|
S01-1028 |
all the features in the Decision
|
List algorithm
|
, expecting that the most informative
|
W02-1003 |
problems that probabilistic decision
|
list algorithms
|
have been used for are very similar
|
P95-1026 |
supervised training using the decision
|
list algorithm
|
, applied to the same data and
|
S01-1028 |
, expecting that the Decision
|
List algorithm
|
would be powerful enough to choose
|
S01-1028 |
BCU-dlist-ehu-all We trained our Decision
|
List algorithm
|
using local and global features
|
P94-1013 |
Yarowsky , 1994 ) , the decision
|
list algorithm
|
outperformed both an N-Gram tagger
|
W02-1027 |
" ) . To compare our decision
|
list algorithm
|
role - for-verb + role to a selectional
|
D09-1048 |
e.g. , semisupervised decision
|
list algorithm
|
( Yarowsky , 1995 ) and Hyperlex
|