Skip to main content
. Author manuscript; available in PMC: 2021 Jul 29.
Published in final edited form as: Comput Linguist Assoc Comput Linguist. 2015 Dec 1;41(4):549–578. doi: 10.1162/coli_a_00232

Table 8.

Classification accuracy results measured by AUC (standard deviation).

Model Summary scores Element scores Element subset
Berkeley-Small 73.7 (3.74) 77.9 (3.52) 80.3 (3.4)
Berkeley-Big 75.1 (3.67) 79.2 (3.45) 81.2 (3.3)
Graph-Small 74.2 (3.71) 78.9 (3.47) 80.0 (3.4)
Graph-Big 74.8 (3.69) 78.6 (3.49) 81.6 (3.3)
Manual Scores 73.3 (3.76) 81.3 (3.32) 82.1 (3.3)
MMSE 72.3 (3.8) n/a n/a
LSA 74.8 (3.7) n/a n/a
BLEU 73.6 (3.8) n/a n/a
ROUGE-SU4 76.6 (3.6) n/a n/a
Unigram overlap precision 73.3 (3.8) n/a n/a
Exact match open-class 74.3 (3.7) 76.4 (3.6) n/a