| N09-2066 | to a better accuracy than the | MST algorithm | . The somewhat surprising result |
| D12-1028 | However , the general non-projective | MST algorithm | enable non-projective edges , |
| D10-1125 | can be found efficiently using | MST algorithms | ( McDonald et al. , 2005 ) . |
| D10-1125 | existing dynamic programming and | MST algorithms | . There are a number of possible |
| P14-1003 | algorithm , the Eisner algorithm and | MST algorithm | are used to parse the test documents |
| P13-1028 | is important to note that the | MST algorithm | may produce non-projective trees |
| P14-1003 | . The Eisner algorithm and the | MST algorithm | are applied to parse the optimal |
| P14-1003 | non-projective dependencies produced by the | MST algorithm | are even reasonable than what |
| W06-2934 | Decoding Instead of using the | MST algorithm | ( McDonald et al. , 2005b ) to |
| P14-1003 | i.e. the Eisner algorithm and | MST algorithm | , are presented to parse the |
| D11-1138 | use the non-projective k-best | MST algorithm | to generate k-best lists ( Hall |
| P14-1003 | Es - pecially , when applying | MST algorithm | on discourse parsing , unlabeled |
| W06-2936 | we will be forced to modify the | MST algorithm | in some ways . 5.2 Swedish Due |
| Q13-1004 | parsing , as well as the k best | MST algorithm | ( Hall , 2007 ) to find the k |
| P14-2106 | using the Chu-Liu/Edmonds directed | MST algorithm | . After several tests we noticed |
| W06-2934 | , even when using the original | MST algorithm | . 4.1 Chinese For Chinese the |
| P14-1003 | labeled accuracy around 0.26 , while | MST algorithm | achieves unlabeled accuracy around |
| P09-1041 | tree constraint imposed by the | MST algorithm | helps information propagate across |
| N06-2033 | using the Chu-Liu/Edmonds directed | MST algorithm | ( Chu & Liu , 1965 ; Edmonds |
| P14-1003 | trees and somewhat suppresses the | MST algorithm | to exhibit its advantage of parsing |
