J. Sarkar, average size 4); and by 10.9 points for S-ALL (sentences from minimum size 1 to maximum size 134 with the average size 22) The suspicion tree built after the Final step shows that coordination cases appear as most suspicious forms Although the corrections made for coordination in the grammar G(Coord) permit generating simple coordinations (e.g., John and Mary likes beans., We played on the roof and in the garden., I cooked beans and she ate it.), the grammar still fails to generate for more complex coordination phenomenon (e.g., verb coordination I cooked and ate beans., gapping phenomenon John eat fish and Harry chips., I liked beans that Harry cooked and which Mary ate Other top suspicious forms are multiword expressions (e.g., at least, so far) and foreign words (part-of-speech FW) (e.g., the naczelnik, perestroika, by 10.4 points for NP-ALL (NPs from minimum size 1 to maximum size 91 with the, 1996.

A. Belz, M. White, D. Espinosa, E. Kow, D. Hogan et al., The first surface realisation shared task: Overview and evaluation results, Proceedings of the 13th European Workshop on Natural Language Generation (ENLG), 2011.

A. Boyd, M. Dickinson, and D. Meurers, On Detecting Errors in Dependency Treebanks, Research on Language and Computation, vol.90, issue.2, pp.113-137, 2008.
DOI : 10.1007/s11168-008-9051-9

Y. Chi, Y. Yang, and R. R. Muntz, Hybridtreeminer: An efficient algorithm for mining frequent rooted trees and free trees using canonical form, Proceedings of the 16th International Conference on and Statistical Database Management (SSDBM), pp.11-20, 2004.

D. De-kok, J. Ma, and G. Van-noord, A generalized method for iterative error mining in parsing results, Proceedings of the 2009 Workshop on Grammar Engineering Across Frameworks, GEAF '09, pp.71-79, 2009.
DOI : 10.3115/1690359.1690368

M. Dickinson and A. Smith, Detecting dependency parse errors with minimal resources, Proceedings of the 12th International Conference on Parsing Technologies, 2011.

C. Gardent and S. Narayan, Error mining on dependency trees, Proceedings of the 50th Meeting of the Association for Computational Linguistics (ACL), pp.592-600, 2012.
URL : https://hal.archives-ouvertes.fr/hal-00768204

S. Narayan and C. Gardent, Structure-driven lexicalist generation, Proceedings of the 24th International Conference on Computational Linguistics (COLING), 2012.
URL : https://hal.archives-ouvertes.fr/hal-00768228

J. R. Quinlan, Induction of decision trees, Machine Learning, pp.81-106, 1986.
DOI : 10.1007/BF00116251

R. Rajkumar, D. Espinosa, and M. White, The osu system for surface realization at generation challenges 2011, Proceedings of the 13th European Workshop on Natural Language Generation (ENLG), pp.236-238, 2011.

B. Sagot and E. De-la-clergerie, Error mining in parsing results, Proceedings of the 21st International Conference on Computational Linguistics and the 44th annual meeting of the ACL , ACL '06, pp.329-336, 2006.
DOI : 10.3115/1220175.1220217

A. Sarkar and A. K. Joshi, Coordination in Tree Adjoining Grammars, Proceedings of the 16th conference on Computational linguistics -, pp.610-615, 1996.
DOI : 10.3115/993268.993275

G. Van-noord, Error mining for wide-coverage grammar engineering, Proceedings of the 42nd Annual Meeting on Association for Computational Linguistics , ACL '04, pp.446-453, 2004.
DOI : 10.3115/1218955.1219012