P11-1151 of the three feature models for local classification . All accuracies are given as
J08-3005 constraints can significantly aid local classification . Because the most dominant constraint
D14-1033 level special module against the local classification approach . The special module
D15-1018 baseline ( SVM ) uses the SVM local classification results from Section 4.2 . The
N07-1047 supervised HMM method embedded with local classification to find the most likely sequence
N07-1047 HMM method in conjunction with a local classification model to predict a global phoneme
P09-2064 incorporate the relation information to local classification results , we employ re-ranking
P09-2064 thematic hierarchy relations into local classification results using re-ranking technique
P09-2064 be viewed as modification for local classification results with structural information
P07-1096 the order of inference and the local classification are dynamically incorporated
P05-1020 Coreference Resolution : From Local Classification to Global Ranking </title> Vincent
D09-1018 classifiers attempt to improve upon the local classifications , Local is also a baseline for
N07-3002 through the use of conventional , local classification methods . In particular , I show
P07-1020 approach , we formulate a sequential local classification problem as shown in Equations
P05-1073 and 59.4 Frame Accuracy ; the local classification model achieves 92.3 F-Measure
N07-3002 improve the training of standard local classification methods , in the context of structured
D14-1033 dependency tree directly , while local classification approaches make predictions on
H05-1059 output by SVMs as the confidence of local classification . <title> Proceedings of Human
N07-1030 partitions are all instances of the local classification approach , and they all use very
J08-3005 also significantly improves the local classification task . Specifically , we show
hide detail